EO4SD-Urban Project: Bamako City Report...Development – Urban Project Issue/Rev-No.: 2.0...

mwanza

Earth Observation for Sustainable Development

Urban Development Project

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 685761.

ESA Ref: AO/1-8346/15/I-NB

Doc. No.: City Operations Report

Issue/Rev.: 2.0

Date: 19/03/2019

EO4SD-Urban Project: Bamako City Report

Lead: Partners: Financed by:

Earth Observation for Sustainable Doc. No.: City-Operations Report

Development – Urban Project Issue/Rev-No.: 2.0

EO4SD-Urban Bamako City Operations Report Page I

Consortium Partners

No. Name Short Name Country

1 GAF AG GAF Germany

2 Système d'Information à Référence Spatiale SAS SIRS France

3 GISAT S.R.O. GISAT Czech Republic

4 Egis SA EGIS France

5 Deutsche Luft- und Raumfahrt e. V DLR Germany

6 Netherlands Geomatics & Earth Observation B.V. NEO The Netherlands

7 JOANNEUM Research Forschungsgesellschaft mbH JR Austria

8 GISBOX SRL GISBOX Romania

Disclaimer:

The contents of this document are the copyright of GAF AG and Partners. It is released by GAF AG

on the condition that it will not be copied in whole, in section or otherwise reproduced (whether by

photographic, reprographic or any other method) and that the contents thereof shall not be divulged to

any other person other than of the addressed (save to the other authorised officers of their organisation

having a need to know such contents, for the purpose of which disclosure is made by GAF AG)

without prior consent of GAF AG.



EO4SD-Urban Bamako City Operations Report Page II

Summary

This document contains information related to the provision of geo-spatial products from the European

Space Agency (ESA) supported project “Earth Observation for Sustainable Development” Urban Applications (EO4SD-Urban) to the World Bank Urban Planning Study for Mali programme for the

City of Bamako.

Affiliation/Function Name Date

Prepared GISAT S.R.O. Václav Stonáček, Jan Kolomazník,

Erika Orlitová

06/03/2019

Reviewed GAF AG Amelie Broszeit 14/03/2019

Approved GAF AG, Project Coordinator Thomas Häusler 14/03/2019

The document is accepted under the assumption that all verification activities were carried out

correctly and any discrepancies are documented properly.

Distribution

Affiliation Name Copies

ESA Z. Bartalis electronic copy

World Bank Jon Kher Kaw, Hyunji Lee electronic copy

Document Status Sheet

Issue Date Details

1.0 29/03/2019 First Document Issue

2.0 04/10/2019 Second Document Issue

Document Change Record

# Date Request Location Details



EO4SD-Urban Bamako City Operations Report Page III



EO4SD-Urban Bamako City Operations Report Page IV

Executive Summary

The European Space Agency (ESA) has been working closely together with the International Finance

Institutes (IFIs) and their client countries to demonstrate the benefits of Earth Observation (EO) in the

IFI development programmes. Earth Observation for Sustainable Development (EO4SD) is an ESA

initiative, which aims to achieve an increase in the uptake of satellite based information in the regional

and global IFI programmes. The overall aim of the EO4SD Urban project is to integrate the

application

of satellite data for urban development programmes being implemented by the IFIs or Multi-Lateral

Development Banks (MDBs) with the developing countries. The overall goal will be achieved via

implementation of the following main objectives:

To provide a service portfolio of Baseline and Derived urban-related geo-spatial products

To provide the geo-spatial products and services on a geographical regional basis

To ensure that the products and services are user-driven

The report describes the methodological approach to produce geospatial products and results of urban

analytics derived from the products as implemented as a part of the ESA funded EO4SD Urban project

for Bamako, Mali in collaboration with the World Bank. EO-based information support for Bamako

has been provided to two teams/activities:

World Bank Urban GP supported UrbanScapes Community-of-Practice (CoP) platform under

an umbrella of ASA programme: “Transforming Cities through Public Spaces”

World Bank supported project “Engine of Growth and Service Delivery” in Mali.

The Report provides a Service Description by referring to the user driven service requirements and the

associated product list with the detailed product specifications. The following products were requested

and delivered:

Settlement Extent and Imperviousness

Urban Land Use / Land Cover

Urban Extent

Urban Open and Green Areas (optional product)

Transport Infrastructure – Road Network

This City Operations Report for Bamako systematically reviews the main production steps involved

and importantly highlights the Quality Control (QC) mechanisms involved; the steps of QC and the

assessment of quality is provided in related QC forms in the Annexe of this Report. There is also the

provision of standard analytical work undertaken with the products, which can be further included as

inputs into further urban development assessments, modelling and reports. The text of the Report is

accompanied by several maps, charts and tables with statistics. Total built-up extent and level of

imperviousness as of 2015 are estimated for the extended (peri-urban) area. LU/LC distribution and its

structure relevant for 2018 are provided for the core urban area of interest, as well as distribution and

typology of road network and structure of urban open and green areas. The last topic is supplemented

by several additional analytics aimed at assisting assessment of public spaces represented by subset of

Open and Green Areas product.



EO4SD-Urban Bamako City Operations Report Page V

This page is intentionally left blank!



EO4SD-Urban Bamako City Operations Report Page VI

Table of Contents

1 GENERAL BACKGROUND OF EO4SD-URBAN ................................................................... 1

2 SERVICE DESCRIPTION........................................................................................................... 1

2.1 STAKEHOLDERS AND REQUIREMENTS ..................................................................................... 1

2.2 SERVICE AREA SPECIFICATION ............................................................................................... 1

2.3 PRODUCT LIST AND PRODUCT SPECIFICATIONS ...................................................................... 2

2.4 LAND USE/LAND COVER NOMENCLATURE ............................................................................. 3

2.5 SETTLEMENT EXTENT .............................................................................................................. 5

2.6 PERCENTAGE IMPERVIOUS SURFACE ...................................................................................... 5

2.7 TRANSPORT INFRASTRUCTURE NOMENCLATURE ................................................................... 6

2.8 OPEN AND GREEN AREAS ........................................................................................................ 7

2.9 TERMS OF ACCESS ................................................................................................................... 7

3 SERVICE OPERATIONS ............................................................................................................ 8

3.1 SOURCE DATA ......................................................................................................................... 8

3.2 PROCESSING METHODS ........................................................................................................... 8

3.3 ACCURACY ASSESSMENT OF MAP PRODUCTS ........................................................................ 9

3.3.1 The Accuracy Assessment of the LU/LC Product ............................................................................. 9

3.3.2 The Accuracy Assessment of the Settlement Extent Product .......................................................... 12

3.3.3 The Accuracy Assessment of the Percentage Impervious Surface Product .................................... 15

3.3.4 The Accuracy Assessment of the Transport Network ..................................................................... 16

3.3.5 The Accuracy Assessment of the Open and Green Areas ............................................................... 18

3.4 QUALITY CONTROL/ASSURANCE .......................................................................................... 20

3.5 METADATA ............................................................................................................................ 21

4 ANALYSIS OF MAPPING RESULTS ..................................................................................... 22

4.1 SETTLEMENT EXTENT – DEVELOPMENTS 1985, 1990, 1995, 2000, 2005, 2010 AND 2015 ... 22

4.2 LAND COVER LAND USE 2018 .............................................................................................. 25

4.3 TRANSPORT NETWORK .......................................................................................................... 28

4.4 URBAN GREEN AREAS AND OPEN AREAS ............................................................................. 30

4.5 CONCLUDING POINTS ............................................................................................................ 36

5 REFERENCES ............................................................................................................................ 37

Annexes

Annex 1: Processing Methods for EO4SD-Urban Products

Annex 2: Filled Quality Control Sheets



EO4SD-Urban Bamako City Operations Report Page VII

List of Figures

Figure 1: Illustration of Core Area of Mapping for Bamako. ............................................................. 2

Figure 2: Mapping result of the city of Bamako of the year 2018 overlaid with randomly distributed

sample points used for accuracy assessment. ........................................................................................ 11

Figure 3: Example of the applied sampling design to generate randomly distributed point for the

Accuracy Assessment of the road network. .......................................................................................... 17

Figure 4: Secondary sampling grid to generate the sampling points at spatial intersection of roads

and grid cells. Roads are represented as white lines, grid as black lines and final sampling point as

black dots. 17

Figure 5: Result of the Urban Green Area mapping in the city of Bamako for the year 2018 overlaid

with randomly distributed sample points used for accuracy assessment. .............................................. 19

Figure 6: Quality Control process for EO4SD-Urban product generation. At each intermediate

processing step output properties are compared against pre-defined requirements. ............................. 20

Figure 7: Urban Extent Status (left) and Urban Imperviousness (right) for Bamako - 2015. ........... 23

Figure 8: Settlement Extent developments between the years 1985 and 2015 in Bamako within

the Core Urban Area .............................................................................................................................. 23

Figure 9: Settlement Imperviousness developments between the years 2005 (left) and 2015

(right) in Bamako within the Core Urban Area ..................................................................................... 24

Figure 10: Detailed Land Cover Land Use 2018 in Bamako .......................................................... 25

Figure 11: Detailed Land Cover Land Use 2018 structure: Presented for Bamako in % (above) and

km2 (below) 26

Figure 12: Transport Network of Bamako in 2018 ............................................................................... 28

Figure 13: Density of street intersections: Number of intersections per 1 km2 of urban fabric per ward

(left) and intersections per 1 km2 grid (right) ........................................................................................ 29

Figure 13: Map of Open and Green Areas based on extended nomenclature – overview (left) and detail

(right) maps – Bamako 2018 ................................................................................................................. 30

Figure 14: Share of OGA: Total area of Open and Green Areas in wards (ha) – left; Share of Open and

Green Areas on Urban fabric per ward (%) – right, Bamako 2018 ....................................................... 31

Figure 15: Open Green Areas Connectivity: Mean distance between nearest Open Green Areas in

wards (m) – left; Euclidian distance to the nearest Open Green Areas calculated for each Open Green

Space – right; Bamako 2018 ................................................................................................................. 32

Figure 16: Open Green Areas Accessibility: Street path distance to the nearest Open Green Space (m)

– left; Average street path distance to the nearest Open Green Space (m) per ward – right; Bamako

2018 ....................................................................................................................................................... 33

Figure 17: Directional distribution and density (as % of Urban Fabric) of OGA classes within regular

distance bands the CBD in Bamako ...................................................................................................... 34

Figure 18: Structure of OGA as a function of distance from the CBD in Bamako ............................... 35

Figure 19: Level of urbanity vs. public space size for two classes (cut-off at 99% percentile of public

space size) ............................................................................................................................................. 36

List of Tables

Table 1: LU/LC Nomenclature for 2018. .......................................................................................... 4

Table 2: Number of sampling points for the EO4SD-Urban mapping classes after applied sampling

design with information on overall land cover by class for year 2018 .................................................. 10

Table 5: Accuracies exhibited by the WSF2015 according to the three considered agreement

criteria for different definitions of settlement. ...................................................................................... 14

Table 4: Acquisition dates and size of the WV2 images available for the 5 test sites analysed in the

validation exercise along with the number of corresponding 30x30m validation samples. .................. 16

Table 5: Validation results of the complemented Transport Network in Bamako, which is based on

OSM data, for year 2018. ...................................................................................................................... 18

Table 6: Detailed information on area and percentage of total area for each class in Bamako. ...... 27

Table 9: Inclusivity: Percentage of population living within 400m catchment areas recalculated for

year 2018 ............................................................................................................................................... 32



EO4SD-Urban Bamako City Operations Report Page VIII

List of Abbreviations CDS City Development Strategy

CS Client States

DEM Digital Elevation Model

DLR German Space Agency

EEA European Environmental Agency

EGIS Consulting Company for Environmental Impact Assessment and Urban Planning, France

EO Earth Observation

ESA European Space Agency

EU European Union

GAF GAF AG, Geospatial Service Provider, Germany

GIS Geographic Information System

GISAT Geospatial Service Provider, Czech Republic

GISBOX Romanian company with activities of Photogrammetry and GIS

GUF Global Urban Footprint

HR High Resolution

HRL High Resolution Layer

IFI International Financing Institute

INSPIRE Infrastructure for Spatial Information in the European Community

ISO/TC 211 Standardization of Digital Geographic Information

JR JOANNEUM Research, Austria

LULC Land Use / Land Cover

LULCC Land Use and Land Cover Change

MMU Minimum Mapping Unit

NDVI Normalized Difference Vegetation Index

NEO Geospatial Service Provider, The Netherlands

OGA

QA

Open and Green Areas

Quality Assurance

QC Quality Control

QM Quality Management

SP Service Provider

VHR Very High Resolution

WB World Bank

WBG World Bank Group



EO4SD-Urban Bamako City Operations Report Page 1

1 General Background of EO4SD-Urban

Since 2008 the European Space Agency (ESA) has worked closely together with the International

Finance Institutes (IFIs) and their client countries to harness the benefits of Earth Observation (EO) in

their operations and resources management. Earth Observation for Sustainable Development (EO4SD)

is a new ESA initiative, which aims to achieve an increase in the uptake of satellite based information

in the regional and global IFI programmes. The EO4SD-Urban project initiated in May 2016 (with a

duration of 3 years) has the overall aim to integrate the application of satellite data for urban

development programmes being implemented by the IFIs with the developing countries. The overall

goal will be achieved via implementation of the following main objectives:

To provide the services on a regional basis (i.e. large geographical areas); in the context of the

current proposal with a focus on S. Asia, SE Asia and Africa, for at least 35-40 cities.

To ensure that the products and services are user-driven; i.e. priority products and services to

be agreed on with the MDBs in relation to their regional programs and furthermore to

implement

the project with a strong stakeholder engagement especially in context with the validation of

the

products/services on their utility.

To provide a service portfolio of Baseline and Derived urban-related geo-spatial products that

have clear technical specifications, and are produced on an operational manner that are

stringently quality controlled and validated by the user community.

To provide a technology transfer component in the project via capacity building exercises in

the

different regions in close co-operation with the MDB programmes.

This Report supports the fulfilment of the third objective which requires the provision of geo-spatial

Baseline and Derived geo-spatial products to various stakeholders in the IFIs and counterpart City

Authorities. The Report provides a service description, and then in Chapter 3 systematically reviews

the main production steps involved and importantly highlights whenever there are Quality Control

(QC) mechanisms involved with the related QC forms in the Annexe of this Report. The description of

the processes is kept intentionally at a top level and avoid technical details as the Report is considered

mainly for non-technical IFI staff and experts and City Authorities. Finally Chapter 4 presents the

standard analytical work undertaken with the products which can be an inputs into further urban

development assessments, modelling and reports.

2 Service Description

The following Sections summarise the service as it has been realised for the core area of the city of

Bamako, Mali within the EO4SD-Urban Project and as it had been delivered to the World Bank’s UrbanScapes team in March 2019.

2.1 Stakeholders and Requirements

EO-based information support for Bamako has been provided to the World Bank Urban GP’s Advisory Services and Analytics (ASA) Programme on “Urban Spaces for City Transformation”, to World Bank supported project “Engine of Growth and Service Delivery” in Mali and to the

counterpart City Authorities in Bamako.

Requirements for Open & Green Spaces analytics were collected from UrbanScapes Community-of-

Practice (CoP) platform under WB‘s ASA programme, whose objective is to develop a framework to examine how urban spaces can transform urban environments by promoting inclusive green growth

and enhancing livability in cities. Specifically, UrbanScapes required development of enhanced

diagnostics based on spatial analysis of localization and characteristics of public spaces to facilitate:.




- Prioritization of interventions improving livability;

- Evidence-based discussion with local city partners;

- City-wide and inter-city assessment of public spaces patterns.

Main objective of support to Engine of Growth project was to provide detailed inventory of public

spaces derived from satellite imagery in order to facilitate assessment of spatial dimensions related to

the project’s operations and development challenges in improving public spaces in Bamako. The

Bamako team was supposed to follow up with ground truthing and local surveys to gather information

including ownership of land in the next steps.

2.2 Service Area Specification

The Areas of Interest (AOI) for mapping the Urban Area for Bamako was depicted in a power point

slide, and sent to the Users for verification. The boundary depicted is based on the municipality and

administration boundaries of the cities. These boundaries were obtained from the GADM database of

Global Administrative Areas (http://www.gadm.org/).

In addition, the AOI area was adjusted based on the population distribution data from 2015 from

WorldPop (https://www.worldpop.org/) and on visual interpretation of the built-up areas as evidenced

on Google Earth. WorldPop has currently the finest resolution global population distribution data

(~100m spatial resolution) available and represents the ambient population.

Figure 1: Illustration of Core Area of Mapping for Bamako.

The Core region has an area of 250 km2.

2.3 Product List and Product Specifications

During the discussions related to the AOIs the potential geo-spatial products that could be provided for

the Cities were also reviewed with the WB Team and Users. It was noted that the Baseline Land

Use/Land Cover (LU/LC) product (for the Core area) is a standard product that would be provided for

http://www.gadm.org/




all Cities as it is required for the derived products. In the case of Bamako, the full list of products for

the Core areas are as follows:

Settlement Extent and Imperviousness

Urban Land Use / Land Cover

Urban Extent

Urban Green Areas

Transport Infrastructure – Road Network

2.4 Land Use/Land Cover Nomenclature

A pre-cursor to starting production was the establishment with the stakeholders on the relevant Land

Use/Land Cover (LU/LC) nomenclature as well as class definitions. The approach taken was to use a

standard remote sensing based LU/LC nomenclature and then adapt it to the User’s LU/LC

requirements. Thus the remote-sensing based LU/LC classes in the urban context can be grouped into

5 Level 1 classes, which are Artificial Areas, Natural/ Semi Natural, Agricultural, Wetland and Water

bodies. These classes can then be sub-divided into several different more detailed classes such that the

dis-aggregation can get down to Level 2-4. This hierarchical classification system is often used in

operational urban mapping programmes and is the basis for example of the European Commission’s Urban Atlas programme which provides pan-European comparable LU/LC data with regular updates.

A depiction of the way the levels and classes are structured is presented as follows:

Level I Artificial Surfaces

- Level II Urban Fabric

Level III

Continuous Urban Fabric (Sealing Layer-S.L. > 80%)

Discontinuous Urban Fabric (S.L. 10% - 80%)

Discontinuous Dense Urban Fabric (S.L. 50% - 80%)

Discontinuous Medium Density Urban Fabric (S.L. 30% - 50%)

Discontinuous Low Density Urban Fabric (S.L. 10% - 30%)

Discontinuous Very Low Density Urban Fabric (S.L. < 10%)

- Level II

Industrial, commercial, public, military, private and transport units

Level III

Industrial, commercial, public, military and private units zoning data

Road and rail network and associated land (Open Street Map or In-country data needed)

o Level IV Fast transit roads and associated land

(Reference: European Union, 2011)

For the Core Urban areas using Very High Resolution (VHR) data it is possible to go down to Level

III and IV. The different levels, classes and sub-classes from the remote sensing based urban

classification, were adapted to the User requirements based on existing Master Plans for cities and/or

direct advice from the User on critical classes required. The final LU/LC nomenclature had to be

endorsed by the User before production started.

In order to link the Urban Atlas classes described in the previous Section with the city of Bamako, the

Consortium used the documentation provided by the Cities for their Master Plans, and assessed which

LU/LC classes could be mapped with remote sensing and linked to the Urban Atlas nomenclature. The

merging of LU/LC classes was provided to the WB Team and Users for review and endorsement. See

Table 1 for the final LU/LC nomenclature used for the year 2018.




Table 1: LU/LC Nomenclature for 2018.

2018

Level I Level II Level III Level IV

1000

Artificial Surfaces

1100

Residential

1100 Residential 1110 Very Low Density

1120 Low Density

1130 Medium Density

1140 High Density

1150 Very High Density

1200

Industrial, Commercial, Public,

Military, Private and Transport

Units

1210 Industrial, Commercial,

Public, Military and Private

Units

1211 Commercial

1212 Industry

1213 University

1214 Schools

1215 Government

1216 Military

1217 Hospitals

1218

Public Buildings

1219 Non-Residential Urban Fabric

1220 Roads 1221 Arterial

1222 Collector

1230 Railway

1240 Airport

1250 Port

1300

Mine, Dump and Construction

Sites

1310 Mineral Extraction and

Dump Sites

1320 Construction Sites

1330 Vacant Land not

obviously being prepared for

construction

1340 Destructed

1400

Urban Open Spaces

1410 Urban Parks

1420 Recreation Facilities

(Sport Facilities, Stadiums,

Golf Courses, etc.)

1430 Cemeteries

2000

Agricultural Area

3000

Natural and Semi-

natural Areas

3100

Forest

3200

Natural Areas (Savannah,

Grassland)

3300

Bare Soil

4000 Wetlands

5000 Water




It is important to note that the possibility to classify at Level IV is highly dependent on the availability

of reliable reference datasets from the City or sources such as Google Earth. This aspect is further

discussed in Chapter 3.

2.5 Settlement Extent

Reliably outlining settlements is of high importance since an accurate characterization of their extent

is fundamental for accurately estimating, among others, the population distribution, the use of

resources (e.g. soil, energy, water, and materials), infrastructure and transport needs, socioeconomic

development, human health and food security. Moreover, monitoring the change in the extent of

settlements over time is of great support for properly modelling the temporal evolution of urbanization

and thus, better estimating future trends and implementing suitable planning strategies.

At present, no standard exists for defining settlements and worldwide almost each country applies its

own definition either based on population, administrative or geometrical criteria. When generating the

settlement extent maps from HR imagery, pixels are labelled as settlement if they intersect any

building, lot or – just within urbanized areas – roads and paved surface where we define:

building as any structure having a roof supported by columns or walls and intended for the

shelter, housing, or enclosure of any individual, animal, process, equipment, goods, or

materials of any kind;

lot as the area contained within an enclosure (wall, fence, hedge) surrounding a building or a

group of buildings. In cases where there are many concentric enclosures around a building, the

lot is considered to stop at the inner most enclosure;

road as any long, narrow stretch with a smoothed or paved surface, made for traveling by

motor vehicle, carriage, etc., between two or more points;

paved surface as any level horizontal surface covered with paving material (i.e., asphalt,

concrete, concrete pavers, or bricks but excluding gravel, crushed rock, and similar materials).

Instead, pixels not satisfying this condition are marked as non-settlement.

The settlement extent product is a binary mask outlining - in the given area of interest (AOI) –

settlements in contrast to all other land-cover classes merged together into a single information class.

The settlement class and the non-settlement class are associated with values “255” and “0”, respectively.

2.6 Percentage Impervious Surface

Settlement growth is associated not only to the construction of new buildings, but – more in general –

to a consistent increase of all the impervious surfaces (hence also including roads, parking lots,

squares, pavement, etc.), which do not allow water to penetrate, forcing it to run off. To effectively

map the percentage impervious surface (PIS) is then of high importance being it related to the risk of

urban floods, the urban heat island phenomenon as well as the reduction of ecological productivity.

Moreover, monitoring the change in the PIS over time is of great support for understanding, together

with information about the spatiotemporal settlement extent evolution, also more details about the type

of urbanization occurred (e.g., if areas with sparse buildings have been replaced by highly impervious

densely built-up areas or vice-versa).

In the framework of the EO4SD-Urban project we generate PIS maps where one pixel is associated

with the estimated percentage of the corresponding surface at the ground covered by buildings or

paved surfaces, where we define:

building as any structure having a roof supported by columns or walls and intended for the

shelter, housing, or enclosure of any individual, animal, process, equipment, goods, or materials

of any kind;




paved surface as any level horizontal surface covered with paving material (i.e. asphalt, concrete,

concrete pavers, or bricks but excluding gravel, crushed rock, and similar materials).

The product provides for each pixel in the considered AOI the estimated PIS. Specifically, values are

integer and range from 0 (no impervious surface in the given pixel) to 100 (completely impervious

surface in the given pixel) with step 5.

2.7 Transport Infrastructure Nomenclature

The road hierarchy used in the classification is based on international road classification standards;

this is for example defined by the European Commission (https://ec.europa.eu/transport/road_safety/-

specialist/knowledge/road/designing_for_road_function/road_classification_en).

Roads are divided into three groups-Arterial or through traffic flow routes (in our case Arterial

Roads), distributor road (in our case Collector Roads), and access roads (in our case Local Roads).

The three road types are defined as follows:

Arterial Roads:

Roads with a flow function allow efficient throughput of (long distance) motorized traffic. All

motorways and express roads as well as some urban ring roads have a flow function. The number of

access and exit points is limited. (https://ec.europa.eu/transport/road_safety/specialist/knowledge/-

road/designing_for_road_function/road_classification_en)

Collector Roads:

Roads with an area distributor function allow entering and leaving residential areas, recreational areas,

industrial zones, and rural settlements with scattered destinations. Junctions are for traffic exchange

(allowing changes in direction etc.); road sections between junctions should facilitate traffic in

flowing.(https://ec.europa.eu/transport/road_safety/specialist/knowledge/road/designing_for_road_fun

ction/road_classification_en)

Local Roads:

Roads with an access function allow actual access to properties alongside a road or street. Both

junctions and the road sections between them are for traffic exchange. (https://ec.europa.eu/transport/-

road_safety/specialist/knowledge/road/designing_for_road_function/road_classification_en).

Arterial roads and collector roads were the main focus of the classification. These types of roads were

identified for the entire AoI. Within the geospatial dataset the road features can be identified within

the attribute table. A value of 1 is assigned to the arterial roads and a value of 2 to the collector lines.

Spatial Accuracy:

The Collector roads and Arterial roads are integrated within the LULC mapping by applying a buffer

around the road centre lines of 12.0 m for the arterial roads and 7.5m for collector roads. The 7.5m are

set as the maximum allowable difference of the mapped centre line in comparison to the location in

the VHR imagery.

https://ec.europa.eu/transport/road_safety/specialist/knowledge/road/designing_for_road_function/road_classification_en

https://ec.europa.eu/transport/road_safety/specialist/knowledge/road/designing_for_road_function/road_classification_en




2.8 Open and Green Areas

As opposed to standard EO4SD-Urban’s Urban Green area Pproduct, which is based on a subset of

LULC classes, this mapping provides Open and Green Areas product with extended nomenclature

supplemented by additional attributes related to open and green spaces patterns a contexts. Structure of

attributes provides means to apply tailored nomenclature using custom rule-based approach. Current

implementation is based on specification defined in cooperation with WBG’s Urban Scapes team.

Further details are provided in the Annex I.

2.9 Terms of Access

The Dissemination of the digital data and the Report was undertaken via FTP.




3 Service Operations

The following Sections present all steps of the service operations including the necessary input data,

the processing methods, the accuracy assessment and the Quality Control procedures. Methods are

presented in a top-level and standardised manner for all the EO4SD-Urban City Reports.

3.1 Source Data

This Section presents the remote sensing and ancillary datasets that were used. Different types of data

from several data providers have been acquired. A complete list of source data as well as a quality

assessment is provided in Annex 2.

A summary of the main data used is provided in the following Sections.

Very High Resolution Optical EO Data

The VHR data for the core urban area mapping had to be acquired and purchased through commercial

EO Data Providers such as Airbus Defence and European Space Imaging.

It has to be noted that under the current collaboration project the VHR EO data had to be purchased

under mono-license agreements between GAF AG and the EO Data Providers. If EO data would have

to be distributed to other stakeholders then further licences for multiple users would have to be

purchased.

The following VHR sensor data have been acquired to cover the entire AoI for Bamako:

WorldView-3: o 3 scenes from 14.01.2018

Detailed lists of the used EO data as well as their quality is documented in the attached Quality

Control Sheets in Annex 2.

Ancillary Data

Open Street Map (OSM) data: OSM data is freely available and generated by volunteers

across the globe. The so called crowd sourced data is not always complete, but has for the

most parts of the world valuable spatial information. Data was downloaded to complement the

Transport Network layer and further enhanced. The spatial location of the OSM based streets

was used as a geospatial reference.

Detailed lists of the used ancillary data as well as their quality is documented in the attached Quality

Control Sheets in Annex 2.

3.2 Processing Methods

Data processing starts at an initial stage with quality checks and verification of all incoming data. This

assessment is performed in order to guarantee the correctness of data before geometric or radiometric

pre-processing is continued. These checks follow defined procedures in order to detect anomalies,

artefacts and inconsistencies. Furthermore, all image and statistical data were visualised and

interpreted by operators.

The main techniques and standards used for data analysis, processing and modelling for each product

are described in Annex 1.




3.3 Accuracy Assessment of Map Products

Data and maps derived from remote sensing contain - like any other map - uncertainties which can be

caused by many factors. The components, which might have an influence on the quality of the maps

derived from EO include quality and suitability of satellite data, interoperability of different sensors,

radiometric and geometric processing, cartographic and thematic standards, and image interpretation

procedures, post-processing of the map products and finally the availability and quality of reference

data. However, the accuracy of map products have a major impact on secondary products and its

utility and therefore an accuracy assessment was considered as a critical component of the entire

production and products delivery process. The main goal of the thematic accuracy assessment was to

guarantee the quality of the mapping products with reference to the accuracy thresholds set by the user

requirements.

The applied accuracy assessments were based on the use of reference data, and applying statistical

sampling to deduce estimates of error in the classifications. In order to provide an efficient, reliable

and robust method to implement an accuracy assessment, there are three major components that had to

be defined: the sampling design, which determines the spatial location of the reference data, the

response design that describes how the reference data is obtained and an analyses design that defines

the accuracy estimates. These steps were undertaken in a harmonised manner for the validation of all

the geo-spatial products.

3.3.1 The Accuracy Assessment of the LU/LC Product

Sampling Design

The sampling design specifies the sample size, sample allocation and the reference assessment units

(i.e. pixels or image blocks). Generally, different sampling schemes can be used in collecting

accuracy assessment data including: simple random sampling, systematic sampling, stratified

random sampling, cluster sampling, and stratified systematic unaligned sampling. In the current

project a single stage stratified random sampling based on the method described by Olofson et al

(20131) was applied which used the map product as the basis for stratification. This ensured that all

classes, even very minor ones were included in the sample.

However, in complex LU/LC products with many classes, this usually results in a large number of

strata (one stratum per LU/LC classes), of which some classes cover only very small areas (e.g. sport

fields, cemeteries) and not being adequately represented in the sampling. In order to achieve a

representative sampling for the statistical analyses of the mapping accuracy it was decided to extend

the single stage stratified random sampling. At the first stage the number of required samples was

allocated within each of the Level I strata. In the second stage all Level III classes that were not

covered by the first sampling, were grouped into one new stratum. Within that stratum the same

number of samples was randomly allocated as the Level I strata received. To avoid a clustering of

point samples within classes and to minimise the effect of spatial autocorrelation a minimum

distance in between the sample points was set to be 150 m. The final sample size for each class can

be considered to be as close as possible to the proportion of the area covered by each stratum

considering that the target was to determine the overall accuracy of the entire map.

The total sample size per stratum was determined by the expected standard error and the estimated

error rate based on the following formula which assumes a simple random sampling (i.e. the

stratification is not considered):

1 Olofsson, P., Foody, G. M., Stehman, S. V., & Woodcock, C. E. (2013). Making better use of accuracy data in

land change studies: Estimating accuracy and area and quantifying uncertainty using stratified estimation.

Remote Sensing of Environment, 129, 122–131. doi:10.1016/j.rse.2012.10.031




n = 𝑃∗𝑞(𝐸𝑧)²

n = number of samples per strata / map class

p = expected accuracy

q = 1 – p

E = Level of acceptable (allowable) sample error

Z = z-value (the given level of significance)

Hence, with an expected accuracy of p = 0.85, a 95% confidence level and an acceptable sampling

error of 5%, the minimum sample size is 196. A 10% oversampling was applied to compensate for

stratification inefficiencies and potentially inadequate samples (e.g. in case of cloudy or shady

reference data). For each Level I strata 215 samples have been randomly allocated. Afterwards,

within all classes of Level III (see Table 2) that did not received samples in the first run, additionally

215 samples were randomly drawn across all these classes.

Table 2: Number of sampling points for the EO4SD-Urban mapping classes after applied sampling

design with information on overall land cover by class for year 2018

Class Name Class

ID

No. of

Sampling

Points

Km² coverage

Residential 1100 185 101.30

Industrial, Commercial, ... 1210 74 32.90

Collector and Arterial Roads 1221,

1222 8 5.02

Railway 1223 3 0.35

Port 1230 1 0.01

Airport 1240 10 6.54

Mining, Dump Sites 1310 9 5.54

Construction 1320 20 10.79

Vacant Land 1330 20 7.58

Urban Green Areas 1410 8 3.19

Sports and Leisure Facilities 1420 12 4.18

Cemeteries 1430 5 1.10

Agriculture 2000 40 28.67

Forest and Shrub Lands 3100 4 1.60

Natural Areas (Grassland) 3200 34 21.83

Bare Soil 3300 10 3.66

Wetlands 4000 4 1.15

Water 5000 13 14.14

Total -- 460 249.55

Response Design

The response design determines the reference information for comparing the map labels to the

reference labels. Collecting reference data on the ground by means of intensive fieldwork is both

costly and time consuming and in most projects not feasible. The most cost effective reference data

sources are VHR satellite data with 0.5 m to 1 m spatial resolution. Czaplewski (2003)2 indicated that

2 Czaplewski, R. L. (2003). Chapter 5: accuracy assessment of maps of forest condition: statistical design and

methodological considerations, pp. 115–140. In Michael A.Wulder, & Steven E. Franklin (Eds.), Remote

sensing of forest environments: concepts and case studies. Boston: Kluwer Academic Publishers (515 pp.).




visual interpretation of EO data is acceptable if the spatial resolution of EO data is sufficiently better

compared to the thematic classification system. However, if there are no EO data with better spatial

resolution available, the assessment results need to be checked against the imagery used in the

production process.

The calculated number of necessary sampling points for each mapping category was randomly

distributed among the strata and overlaid to the VHR data of each epoch. The following Figure is

showing the mapping result with the overlaid sample points.

Figure 2: Mapping result of the city of Bamako of the year 2018 overlaid with randomly distributed

sample points used for accuracy assessment.

In this way a reference information could be extracted for each sample point by visual interpretation of

the VHR data for all mapped classes. The size of the area to be observed had to be related to the

Minimum Mapping Unit (MMU) of the map product to be assessed. The reference information of each

sampling point was compared with the mapping results and the numbers of correctly and not-correctly

classified observations were recorded for each class. From this information the specific error matrices

and statistics were computed (see next Section).

Analysis

Each class usually has errors of both omission and commission, and in most situations, these errors for

a class are not equal. In order to calculate these errors as well as the uncertainties (confidence

intervals) for the area of each class a statistically sound accuracy assessment was implemented.

The confusion matrix is a common and effective way to represent quantitative errors in a categorical

map, especially for maps derived from remote sensing data. The matrices for each assessment epoch

were generated by comparing the “reference” information of the samples with their corresponding classes on the map. The Reference represented the “truth”, while the Map provided the data obtained

from the map result. Thematic accuracy for each class and overall accuracy is then presented in error

matrices (see Tables below). Unequal sampling intensity resulting from the random sampling




approach was accounted for by applying a weight factor (p) to each sample unit based on the ratio

between the number of samples and the size of the stratum considered3: �̂�𝑖𝑗 = ( 1𝑀) ∑ 1𝜋𝑢ℎ∗𝑥∈(𝑖,𝑗)

Where i and j are the columns and rows in the matrix, M is the total number of possible units

(population) and π is the sampling intensity for a given sample unit u in stratum h.

Overall accuracy and User and producer accuracy were computed for all thematic classes and 95%

confidence intervals were calculated for each accuracy metric.

The standard error of the error rate was calculated as follows: 𝜎ℎ = √𝑝ℎ(1−𝑝ℎ)𝑛ℎ where nh is the sample

size for stratum h and ph is the expected error rate. The standard error was calculated for each stratum

and an overall standard error was calculated based on the following formula: 𝜎 = √∑ 𝑤ℎ2. 𝜎ℎ2

In which 𝑤ℎ is the proportion of the total area covered by each stratum. The 95% Confidence Interval

(CI) is +/- 1.96*𝜎.

Results

The confusion matrices are provided within the Annex 2 and showing the mapping error for each

relevant class. For each class the number of samples which are correctly and not correctly classified

are listed, which allows the calculation of the user and producer accuracies for each class as well as

the confidence interval at 95% confidence levels based on the formulae above.

The Land Use/Land Cover product for Bamako has an overall mapping accuracy of 93% with a

CI ranging from 90.7% to 95.4% at a 95% CI. The specific class accuracies are given in Annex

2.

3.3.2 The Accuracy Assessment of the Settlement Extent Product

In the following, we present the strategy designed for validating the World Settlement Footprint

(WSF) 2015, i.e. a global settlement extent layer obtained as a mosaic of ~18.000 tiles of 1x1 degree

size where the same technique employed in the EO4SD-Urban project has been used. In particular,

specific details are given for all protocols adopted for each of the accuracy assessment components,

namely response design, sampling design, and analysis; final results are discussed afterwards. In the

light of the quality and amount of validation points considered, we reasonably assume that the

corresponding quality assessment figures are also representative for any settlement extent map

generated in the framework of EO4SD-Urban.

Response Design

The response design encompasses all steps of the protocol that lead to a decision regarding agreement

of the reference and map classifications. The four major features of the response design are the source

of information used to determine the source of reference data, the spatial unit, the labelling protocol

for the reference classification, and a definition of agreement.

3 Selkowitz, D. J., & Stehman, S. V. (2011). Thematic accuracy of the National Land Cover Database (NLCD)

2001 land cover for Alaska. Remote Sensing of Environment, 115(6), 1401–1407.

doi:10.1016/j.rse.2011.01.020.




Source of Reference Data: Google Earth (GE) satellite/aerial VHR imagery has been used given

its free access and the availability for all the project test sites in the period 2014-2015. In

particular, GE automatically displays the latest available data, but it allows to browse in time over

all past historical images. The spatial resolution varies depending on the specific data source; in

the case of SPOT imagery it is ~1.5m, for Digital Globe's WorldView-1/2 series, GeoEye-1, and

Airbus' Pleiades it is in the order of ~0.5m resolution, whereas for airborne data (mostly available

for North America, Europe and Japan) it is about 0.15m.

Spatial Assessment Unit: A 3x3 block spatial assessment unit composed of 9 cells of 10x10m

size has been used. Specifically, this choice is justified one the one hand by the fact that input

data with different spatial resolutions have been used to generate the WSF2015 (i.e. 30m Landsat-

8 and 10m S1). On the other hand, GE imagery exhibited in some cases a misregistration error of

the order of 10-15m, hence using a 3x3 block allows defining an agreement e.g. based on

statistics computed over 9 pixels, thus reducing the impact of such shift.

Reference Labelling Protocol: For each spatial assessment block any cell is finally labelled as

settlement if it intersects any building, lot or – just within settlements – roads and paved surface.

Instead, pixels not satisfying this condition are marked as non-settlement.

Definition of Agreement: Given the classification and the reference labels derived as described

above, three different agreement criteria have been defined:

1) for each pixel, positive agreement occurs only for matching labels between the

classification and the reference;

2) for each block, a majority rule is applied over the corresponding 9 pixels of both the

classification and the reference; if the final labels match, then the agreement is positive;

3) for the classification a majority rule is applied over each assessment block, while for the

reference each block is labelled as “settlement” only in the case it contains at least one pixel marked as “settlement”; if the final labels match, then the agreement is positive.

Crowd-sourcing was performed internally at Google. In particular, by means of an ad-hoc tool,

operators have been iteratively prompted a given cell on top of the available Google Earth reference

VHR scene closest in time to the year 2015 and given the possibility of assigning to each cell a label

among: “building”, “lot”, “road/paved surface” and “other”. For training the operators, a representative set of 100 reference grids was prepared in collaboration between Google and DLR.

Sampling Design

The stratified random sampling design has been applied since it satisfies the basic accuracy assessment

objectives and most of the desirable design criteria. In particular, stratified random sampling is a

probability sampling design and it is one of the easier to implement; indeed, it involves first the

division of the population into strata within which random sampling is performed afterwards. To

include a representative population of settlement patterns, 50 out of the ~18.000 tiles of 1x1 degree

size considered in the generation of the WSF2015 have been selected based on the ratio between the

number of estimated settlements (i.e. disjoint clusters of pixels categorized as settlement in the

WSF2015) and their area. In particular, the i-th selected tile has been chosen randomly among those

whose ratio belongs to the interval ]𝑃2(𝑖−1); 𝑃2𝑖], 𝑖 ∈ [1; 50] ⊂ ℕ (where 𝑃𝑥 denotes the x-th percentile

of the ratio).

As the settlement class covers a sensibly small proportion of area compared to the merger of all other

non-settlement classes (~1% of Earth’s emerged surface), an equal allocation reduces the standard

error of its class-specific accuracy. Moreover, such an approach allows to best address user’s accuracy estimation, which corresponds to the map “reliability” and is indicative of the probability that a pixel

classified on the map actually represents the corresponding category on the ground. Accordingly, in

this framework for each of the 50 selected tiles we randomly extracted 1000 settlement and 1000 non-

settlement samples from the WSF2015 and used these as centre cells of the 3x3 reference block

assessment units to label by photointerpretation. Such a strategy resulted in an overall amount of (1000 + 1000) × 9 × 50 = 900.000 cells labelled by the crowd.




Analysis

As measures for assessing the accuracy of the settlement extent maps, we considered:

the percentage overall accuracy OA%;

the Kappa coefficient;

the percentage producer’s (PAS%, PANS%) and user’s (UAS%, UANS%) accuracies for both the

settlement and non-settlement class;

the percentage average accuracy AA% (i.e., the average between PAS% and PANS%).

Results

Table 2 reports the accuracies exhibited by the WSF2015 according to the three considered agreement

criteria for different definitions of settlement; specifically, we considered as “settlement” all areas covered by: i) buildings; ii) buildings or building lots; or iii) buildings, building lots or roads / paved

surfaces. As one can notice, accuracies are always particularly high, thus confirming the effectiveness

of the employed approach and the reliability of the final settlement extent maps. The best

performances in terms of kappa are obtained when considering settlements as composed by buildings,

building lots and roads / paved surfaces for criteria 1 and 2 (i.e., 0.6938 and 0.7317, respectively) and

by buildings and building lots for criteria 3 (0.7716); the OA% follows a similar trend. This is in line

with the adopted settlement definition. Moreover, agreement criteria 3 results in accuracies

particularly high with respect to criteria 1 and 2 when considering as settlement just buildings or the

combination of buildings and lots. This can be explained by the fact that when the detection is mainly

driven by Landsat data then the whole 3x3 assessment unit tends to be labelled as settlement if a

building or a lot intersect the corresponding 30m resolution pixel.

Table 3: Accuracies exhibited by the WSF2015 according to the three considered agreement criteria

for different definitions of settlement.

Settlement = Accuracy

Measure

Agreement Criterion

1 2 3

buildings

OA% 86.96 87.86 91.15

AA% 88.57 90.35 88.91

Kappa 0.6071 0.6369 0.7658

UANS% - UAS% 98.11 54.69 98.73 56.76 94.84 80.58

PANS% - PAS% 86.24 90.90 86.72 93.98 93.32 84.51

buildings + lots

OA 88.08 88.94 91.26

AA% 88.64 90.19 88.71

Kappa 0.6510 0.6784 0.7716

UANS% - UAS% 97.54 60.71 98.13 62.66 94.29 82.62

PANS% - PAS% 87.79 89.49 88.26 92.12 93.95 83.48

buildings + lots

+ roads / paved

surface

OA 88.77 90.09 88.51

AA% 86.34 88.28 84.27

Kappa 0.6938 0.7317 0.7219

UANS% - UAS% 94.49 72.20 95.35 75.06 88.13 89.60

PANS% - PAS% 90.78 81.91 91.62 84.94 96.04 72.51




3.3.3 The Accuracy Assessment of the Percentage Impervious Surface

Product

In the following, we present the strategy designed for validating the PIS product; specifically, details

are given for all protocols adopted for each of the accuracy assessment components, namely response

design, sampling design, and analysis. Results are discussed afterwards.

Response Design

The response design encompasses all steps of the protocol that lead to a decision regarding agreement

of the reference and map classifications. The four major features of the response design are the source

of information used to determine the source of reference data, the spatial unit, the labelling protocol

for the reference classification, and a definition of agreement.

Source of Reference Data: Cloud-free VHR multi-spectral imagery (Visible + Near Infrared)

acquired at 2m spatial resolution (or higher) covering a portion of the AOI for which the Landsat-

based PIS product has been generated;

Spatial Assessment Unit: A 30x30m size unit has been chosen according to the spatial resolution

of the Landsat imagery employed to generate the PIS product;

Reference Labelling Protocol: We first compute for each VHR scene the NDVI and manually

identify the most suitable threshold that allows to exclude all the vegetated areas (i.e. non-

impervious). Then, we refine the resulting mask by extensive photointerpretation.

Definition of Agreement: We aggregate the above-mentioned masks at 30m spatial resolution

and compare per-pixel the resulting VHR-based reference PIS to the corresponding portion of the

Landsat-based PIS product.

Sampling Design

The entirety of pixels covered by the available VHR imagery over the given AOI is employed for

assessing the quality of the Landsat-based PIS product.

Analysis

As measures for assessing the accuracy of the PIS maps, we compute:

the Pearson’s Correlation coefficient: it measures the strength of the linear relationship between

two variables and it is defined as the covariance of the two variables divided by the product of

their standard deviations; in particular, it is largely employed in the literature for validating the

output of regression models;

The Mean Error (ME): it is calculated as the difference between the estimated value (i.e., the

Landsat-based PIS) and the reference value (i.e., the VHR-based reference PIS) averaged over all

the pixels of the image;

The Mean Absolute Error (MAE): it is calculated as the absolute difference between the estimated

value (i.e., the Landsat-based PIS) and the reference value (i.e., the VHR-based reference)

averaged over all the pixels of the image.

Results

To assess the effectiveness of the method developed to generate the PIS maps, we analysed its

performances over 5 test sites (i.e. Antwerp, Helsinki, London, Madrid and Milan) by means of

WorldView-2 (WV2) scenes acquired in 2013-2014 at 2m spatial resolution. In particular, given the

spatial detail offered by WV2 imagery, it was possible to delineate with a very high degree of

confidence all the buildings and other impervious surfaces included in the different investigated areas.

Details about acquisition date and size are reported in Table 6, along with the overall number of final

30x30m validation samples derived for the validation exercise. Such a task demanded a lot of manual

interactions and transferring it to other AOIs would require extensive efforts; however, we reasonably

assume that the final quality assessment figures (computed on the basis of more than 1.9 million

validation samples) shall be considered representative also for PIS maps generated in the framework




of EO4SD-Urban. Table 4 reports the quantitative results of the comparison between the PIS maps

generated using Landsat-7/8 data acquired in 2013-2014 and the WV2-based reference PIS maps. In

particular, the considered approach allowed to obtain a mean correlation of 0.8271 and average ME

and MAE equal to -0.09 and 13.33, respectively, hence assessing the great effectiveness of the

Landsat-based PIS products. However, it is worth also pointing out that due to the different acquisition

geometries, WV2 and LS8 images generally exhibit a very small shift. Nevertheless, despite limited,

such displacement often results in a one-pixel shift between the Landsat-based PIS and the WV2-

based reference PIS aggregated at 30m resolution. This somehow affects the computation of the MAE

and of the correlation coefficient (which however yet resulted in highly satisfactory values). Instead,

the bias does not alter the ME which always exhibited values close to 0, thus confirming the

capabilities of the technique and the reliability of the final products.

Table 4: Acquisition dates and size of the WV2 images available for the 5 test sites analysed in the

validation exercise along with the number of corresponding 30x30m validation samples.

3.3.4 The Accuracy Assessment of the Transport Network

The road network was partially integrated in the LU/LC map by selecting first and second level roads.

These are the Arterial Roads and the Collector Lines. For the accuracy assessment of the Road

network it should be noted that the sampling design, response design and analyses design are different

from the one used for validating the LULC maps. The Accuracy Assessment of the Transport Network

is related to the geospatial precision of the collected and digitised centerlines of the roads.

Sampling Methodology

A systematic random sampling was applied to define the primary and secondy sampling units. Over

the entire AoI a regular grid of 450m by 450m was created. Based on these grid cells a random

selection of 2% sample cells were selected. An example is given in Figure 3 with the road network in

grey, the grid cells in black and the randomly selcted cells in green.

Acquisition Date

[DD.MM.YYYY]

Original Size

[2x2m pixel]

Validation Samples

[30x30m unit]

Antwerp 31.07.2014 5404 x 7844 188.280

Helsinki 21.04.2014 12468 x 9323 516.882

London 28.08.2013 7992 x 8832 313.937

Madrid 20.12.2013 10094 x 13105 588.202

Milan 14.05.2014 8418 x 7957 297.330




Figure 3: Example of the applied sampling design to generate randomly distributed point for the

Accuracy Assessment of the road network.

Within the randomly selected cells another grid of 150m distance was created (see Figure 4). All

intersection between the created road layer and the 150m grid were extracted as points. At all points

the road locations were visually checked and if any, the differences between spatial location on VHR

imagery and spatial location of the digitised lines recorded.

Figure 4: Secondary sampling grid to generate the sampling points at spatial intersection of roads and

grid cells. Roads are represented as white lines, grid as black lines and final sampling point as

black dots.

Overall 940 sampling points were created and their differences recorded. The result is presented in

Table 5 as histogram of deviations. For the entire sampling population a Mean Difference of 0.78 m

and a Standard Deviation of 1.26 m was calculated.




Table 5: Validation results of the complemented Transport Network in Bamako, which is based on

OSM data, for year 2018.

Distance in m Frequency

0.0 633

1.0 41

2.0 138

3.0 77

4.0 34

Above 4 meters 17

By setting a maximum allowable difference of 3 m the distances are separated into two classes.

Correct street locations and in-correct street locations. The statistical analysis for the two classes

reveal an overall accuracy of 94.6% was achieved.

3.3.5 The Accuracy Assessment of the Open and Green Areas

Thematic accuracy assessment for Urban Green and Open and Green Areas (OGA) products generally

follows the methodology and protocol as defined by EO4SD-Urban for LULC. Certain specificities

taking into account the complex nature of the OGA product are adopted. Accuracy is assessed step-

wise in stages reflecting the evolution of interim products used to generate final OGA layer. Following

on the LU/LC accuracy assessment 2 additional assessments are carried out with respect to Urban

Greens and OGA.

Sampling design, the first step, is depending on a product either one- or two-stage stratified random

sampling. The two-stage sampling design is used for the standard LU/LC standard product with

several classes while the one-stage sampling design is utilized for specialized optional products such

as OGA. The number of samples is allocated and subsequently randomly distributed to the strata. The

final sample size for each class can be considered to be as close as possible to the proportion of the

area covered by each stratum considering that the target was to determine the overall accuracy of the

entire map (Olofsson, Foody, Stehman, & Woodcock, 2013). Clustering effect and consequent spatial

autocorrelation is avoided by maintaining minimum distance in between the sample points (the default

value is 150 m).

The total sample size per stratum was determined by the expected standard error and the estimated

error rate based on the formula assuming a simple random sampling without considering further

stratification (Olofsson, Foody, Stehman, & Woodcock, 2013).

The response design is based on a pseudo-ground truth using independent interpretation of EO

imagery. A reference information could be extracted for each sample point by visual interpretation

for all mapped classes using the VHR data used in the production process. The size of the area to be

observed is related to the Minimum Mapping Unit (MMU) of the map product. The reference

information of each sampling point is then compared with the mapping results and the numbers of

correctly and not-correctly classified observations are recorded for each class.

Ground truthing needs to be optionally conducted for selected specialized products in cooperation

with task teams, if applicable. This will enable cities to supplement missing components from remote

sensing, such as facilities, user surveys, and ownership surveys. Geographically limited subset of the

sample is expected to be used for in-situ reference data collection. OGA classes will have to be

aggregated prior to the application of the field data collection in order to define strata whose land use

status can be interpreted by field visits.




Analysis design applies an error (confusion) matrix representing means to derive quantitative errors

of a categorical map. Class-wise errors of omission and commission and their 95% confidence level

intervals are denoted, as well as overall accuracy and standard error rates after weighting for strata

area proportions.

1) Accuracy assessment of the Urban Green Areas standard product

Standard Urban Green Areas product is generated as follows: in the first step all vegetated areas are

classified, in the second step the class 1410 and 1430 are added to already classified vegetated areas.

Sampling and response design follow the EO4SD-Urban protocol: samples are allocated within the

two classes and distributed randomly. Sampling only within extent of polygonal subset from LU/LC

would prevent obtaining error of omission. Therefore, additional samples were also collected outside

the polygons. In order to focus on areas with potentially higher probability of being green and to

reduce potential bias the additional samples are dominantly distributed within vacant land and within

segments with vegetation (regardless if artificial or non-artificial) generated by means of automated

classification of reference VHR imagery.

2) Accuracy assessment of the Open and Green Areas optional product

Open and Green Areas product as generated by enhancement of standard Urban Green Areas product:

by supplementing of class 1420 (Sport, leisure and recreational facilities including associated land)

and additional polygons probably related to public function. Two-stage sampling design follows the

one for LU/LC. Classes from Level 2 are aggregated according to their in-situ interpretability to form

logical strata before allocation.

Figure 5: Result of the Urban Green Area mapping in the city of Bamako for the year 2018 overlaid

with randomly distributed sample points used for accuracy assessment.

The Urban Green Area product for Bamako has an overall mapping accuracy of 94.3% with a

CI ranging from 91.7% to 97.0% at a 95% CI. The specific class accuracies are given in Annex

2.




3.4 Quality Control/Assurance

A detailed Quality Control and Quality Assurance (QC/QA) system has been developed which records

and documents all quality relevant processes ranging from the agreed product requirements, the

different types of input data and their quality as well as the subsequent processing and accuracy

assessment steps. The main goal of the QC/QA procedures was the verification of the completeness,

logical consistency, geometric and thematic accuracy and that metadata are following ISO standards

on geographic data quality and INSPIRE data specifications. These assessments were recorded in Data

Quality Sheets which are provided in Annex 2. The QC/QA procedures were based on an assessment

of a series of relevant data elements and processing steps which are part of the categories listed below:

Product requirements;

Specifications of input data: EO data, in-situ data, ancillary data;

Data quality checks: EO data quality, in-situ data quality, ancillary data quality;

Geometric correction, geometric accuracy, data fusion (if applicable), data processing;

Thematic processing: classification, plausibility checks;

Accuracy: thematic accuracy, error matrices

Delivery checks: completeness, compliancy with requirements

After each intermediate processing step a QC/QA was performed to evaluate products appropriateness

for the subsequent processing (see Figure 6).

Figure 6: Quality Control process for EO4SD-Urban product generation. At each intermediate

processing step output properties are compared against pre-defined requirements.

After the initial definition of the product specifications (output) necessary input data were defined and

acquired. Input data include all satellite data and reference data e.g. in-situ data, reference maps,

topographic data, relevant studies, existing standards and specifications, statistics. These input data

were the baseline for the subsequent processing and therefore all input data had to be checked for

completeness, accuracy and consistency. The evaluation of the quality of input data provides

confidence of their suitability for further use (e.g. comparison with actual data) in the subsequent

processing line. Data processing towards the end-product required multiple intermediate processing

steps. To guarantee a traceable and quality assured map production the QC/QA assessment was

performed and documented by personnel responsible for the Quality Control/Assurance. The results of

all relevant steps provided information of the acceptance status of a dataset/product.

The documentation is furthermore important to provide a comprehensive and transparent summary of

each production step and the changes made to the input data. With this information the user will be

able to evaluate the provided services and products. Especially the accuracy assessment of map

products and the related error matrices are highly important to rate the quality and compare map

products from different service providers.

The finalised QC/QA forms are attached in Annex 2.




3.5 Metadata

Metadata provides additional information about the delivered products to enable it to be better

understood. In the current project a harmonised approach to provide metadata in a standardised format

applicable to all products and end-users was adopted. Metadata are provided as XML files, compliant

to the ISO standard 19115 "Metadata" and ISO 19139 "XML Scheme Implementation". The metadata

files have been created and validated by the GIS/IP-operator for each map product with the

Infrastructure for Spatial Information in Europe (INSPIRE) Metadata Editor available at:

http://inspire-geoportal.ec.europa.eu/editor/.

The European Community enacted a Directive in 2007 for the creation of a common geo-data

infrastructure to provide a consistent metadata scheme for geospatial services and products that could

be used not only in Europe but globally. The geospatial infrastructure called INSPIRE was built in a

close relation to existing International Organization for Standardization (ISO) standards. These are

ISO 191115, ISO 19119 and ISO 15836. The primary incentive of INSPIRE is to facilitate the use and

sharing of spatial information by providing key elements and guidelines for the creation of metadata

for geospatial products and services.

The INSPIRE Metadata provides a core set of metadata elements which are part of all the delivered

geo-spatial products to the users. Furthermore, the metadata elements provide elements that are

necessary to perform queries, store and relocate data in an efficient manner. The minimum required

information is specified in the Commission Regulation (EC) No 1205/2008 of 3 December 2008 and

contains 10 elements:

Information on overall Product in terms of: Point of contact for product generation, date of

creation

Identification of Product: Resource title, Abstract (a short description of product) and Locator

Classification of Spatial Data

Keywords (that define the product)

Geographic information: Area Coverage of the Product

Temporal Reference: Temporal extent; date of publication; date of last revision; date of

creation

Quality and Validity: Lineage, spatial resolution

Conformity: degree of conformance to specifications

Data access constraints or Limitations

Responsible party: contact details and role of contact group/person

These elements (not exhaustive) constitute the core information that has to be provided to meet the

minimum requirements for Metadata compliancy. Each element and its sub-categories or elements

have specific definitions; for example in the element “Quality” there is a component called “Lineage” which has a specific definition as follows: “a statement on process history and/or overall quality of the spatial data set. Where appropriate it may include a statement whether the data set has been validated

or quality assured, whether it is the official version (if multiple versions exist), and whether it has legal

validity. The value domain of this element is free text,” (INSPIRE Metadata Technical Guidelines, 2013). The detailed information on the Metadata elements and their definitions can be found in the

“INSPIRE Metadata Implementing Rules: Technical Guidelines,” (2013). Each of the EO4SD-Urban

products will be accompanied by such a descriptive metadata file. It should be noted that the internal

use of metadata in these institutions might not be established at an operational level, but the file format

(*.xml) and the web accessibility of data viewers enable for the full utility of the metadata.

http://inspire-geoportal.ec.europa.eu/editor/




4 Analysis of Mapping Results

This Chapter will present and assess all results which have been produced within the framework of the

current project, in the context of presentation of the Urban Extent product, the LU/LC products and the

Transport Infrastructure product. Furthermore the Sections that follow will provide the results of some

standard analytics undertaken with these products including the following:

Urban Extent – Developments from 1985, 1990, 1995, 2000, 2005, 2010 to 2015

Land Cover Land Use - Status of year 2018

Transport Infrastructure - Status of year 2018

Urban Green Areas – Status of year 2018

It is envisaged that these analytics provide information on general trends and developments in the

Core

areas which can then be further interpreted and used by Urban planners and the City Authorities for

city planning.

It should be noted that all digital data sets for these products are provided in concurrence with this City

Report with all the related metadata and Quality Control documentation

4.1 Settlement Extent – Developments 1985, 1990, 1995, 2000, 2005, 2010

and 2015

The Urban Extent product in the EO4SD-Urban project is provided by the German Aerospace Centre

(DLR) and is provided for 7 points in time; the 2015 World Settlement Footprint (WSF) product has

been produced jointly exploiting multi-temporal 30m Landsat-8 and ESA Sentinel-1 data with 10m

resolution acquired in 2014-2015. And for the years 2000, 2005 and 2010, the Urban Extent products

generated – given the unavailability of freely and easily accessible multi-temporal radar data at high

resolution – were based only on multi-temporal 30m Landsat-5 and Landsat-7 imagery, and scaled up

to 10m resolution.

It should be noted that some structures which are flat in nature such as

airport runways were not classified; this is due to the fact that radar relies on backscatter which is

more prominent from vertical features. The WSF 2015 products have been validated and made

available as public domain data from October 201 7 onwards on the Urban Thematic Exploitation

Platform (TEP) supported by the DLR.

In the current project the Urban Extent product for Bamako was first used to assess historical

developments from 1985-2015. Further analysis by overlaying administrative boundaries can be

performed to assess urbanisation extent patterns based on administrative units.

Results:

The first result provided using the different Urban Extent products from 1985 to 2015 is illustrated in

Figure 7 which shows the urban development in the Core areas as well as surrounding regions of

Bamako. The Urban Extent developments after 2000 can be examined by Urban Planners to identify

different patterns of growth such as “Edge Growth” or “Leapfrog Growth” depending on the location

of the developments.

Figure 7 shows the Settlement Extent and Imperviousness layer as of 2015. In Bamako there are few

parts where the sealing is already very high and other densification is not realistic. Parts of city,

mainly on the edges are still under development so there is a space for further development.




Figure 7: Urban Extent Status (left) and Urban Imperviousness (right) for Bamako - 2015.

Figure 8: Settlement Extent developments between the years 1985 and 2015 in Bamako within the Core

Urban Area

Figure 8 presents the development of settlement extent between years 1985, 1990, 1995, 2000, 2005,

2010 and 2015. There was important expansion axis along the Niger river and in south-west and east

part of the city, outside of the AOI. Traditional villages are or will be gradually attached by growing

city.




Figure 9: Settlement Imperviousness developments between the years 2005 (left) and 2015 (right) in

Bamako within the Core Urban Area

As seen in Figure 9 there is evidence that in the Core city densification occurred in the last decade.

This process is dominant also in the peri-urban area.




4.2 Land Cover Land Use 2018

This Section presents the results of the LU/LC mapping for 2018. The LU/LC overview map for 2018

is depicted in Figure 10 and a cartographic version of the map layout is provided as a pdf file in

addition to the geospatial product.

Further information on the class disaggregation and area coverage is presented in Figure 11 for the

year 2018.

Figure 10: Detailed Land Cover Land Use 2018 in Bamako




Figure 11: Detailed Land Cover Land Use 2018 structure: Presented for Bamako in % (above) and km2

(below)

Bamako is a highly urbanized city with 40.59% of residential urban fabric (different sealing levels)

followed by Industrial, Commercial and other Non-Residential urban fabric (17.96%). Other classes

are quite equally in size. Agriculture area (11.49%) is dominant in south part of AOI between city and

airport. Natural areas (8.75%) are in contrary located in the north-western part of the city. City of

Bamako is divided into two parts by the Niger river.

Detailed information on the area and percentage distribution can be further observed in Table 6.




Table 6: Detailed information on area and percentage of total area for each class in Bamako.

LU/LC Classes 2018

sqkm % of

Residential 0-10 % 0.45 0.18%

Residential 10-30 % 4.46 1.79%

Residential 30-50 % 14.97 6.00%

Residential 50-80 % 58.43 23.51%

Residential 80-100 % 22.98 9.21%

Industrial, Commercial, Public, Military 32.90 17.96%

Arterial Line 0.70 0.28%

Collector Line 4.31 1.73%

Railway 0.35 0.14%

Port 0.01 0.00%

Airport 6.54 2.62%

Mining, Quarry Areas, Dump Sites 5.54 2.22%

Construction Site 10.79 4.32%

Vacant land 7.58 3.04%

Urban Parks 3.19 1.28%

Recreation Facilities 4.18 1.67%

Cemeteries 1.10 0.44%

Agricultural Area 28.67 11.49%

Forest 1.60 0.64%

Natural areas (non-forested) 21.83 8.75%

Bare Soil 3.66 1.47%

Wetlands 1.15 0.46%

Water 14.14 5.67%

Total 249.56 100.00




4.3 Transport Network

The Transport Network was created for current point in time (2018) using three road types. The

Arterial roads and Collector roads were integrated in the LULC map by applying a buffer of 12 m and

8 m for the Arterial and Collector roads, respectively. Local roads are only part of the vector data set,

which are provided to the user.

Figure 12 depicts the Transport Network for the current point in time.

Figure 12: Transport Network of Bamako in 2018




Figure 13: Density of street intersections: Number of intersections per 1 km2 of urban fabric per ward

(left) and intersections per 1 km2 grid (right)

Figure 13 presents Street Intersection Density. Left figure shows number of intersections per 1 km2 of

urban fabric. Intersection is supposed to be a street network node with a minimum of 3 converging

roads (i.e. road dead ends are excluded).

The figure on right presents intersections density per 1km2 visualized in the regular grid.




4.4 Urban Green Areas and Open Areas

This section provides insight into spatial distribution of Green Areas and Open Spaces elements within

the Core city of Bamako: green areas share, greenness index and its evolution, proximity, detailed

situation mapping, fragmentation/connectivity and accessibility.

In addition to standard Urban Green Areas product being part of EO4SD’s products portfolio and consisting of two classes, advanced service with extended detailed typology is provided: Open and

Green Areas (OGA). It was derived by interpretation of VHR imagery as of 2018. Public urban

spaces such as streets, open spaces, green areas, parks, and public buildings are a part of cities that is

often overlooked. Inadequate, poorly designed, or privatized public spaces often lead to exclusion,

marginalization, and degradation of liveability in the urban environment. That is why the importance

of green areas and open spaces are now embedded within the Sustainable Development Goals,

particularly Goal 11.7 on universal access to safe, inclusive and accessible green and public spaces.

The service showcases how the use of geospatial information derived from very high resolution

(VHR) satellite imagery can contribute to development of a framework to examine current situation

and future potential of urban spaces to transform urban environments by promoting inclusive green

growth and enhancing liveability in megacities and, furthermore show how such framework, supported

by satellite image based information, can be scaled up to other global megacities and large cities in

future. Figure 14 shows example of OGA product. More details on Urban Green Areas and Open

Spaces typology and elements identified can be found in Annex 1.

Figure 14: Map of Open and Green Areas based on extended nomenclature – overview (left) and detail

(right) maps – Bamako 2018

Distribution of Open and Green Areas is presented in the Figure 14. Extended typology derived by

CAPI interpretation is demonstrated in detail on the right. This product represents maximalist option

in terms of spatial and thematic detail for open and green areas, which is feasible to be extracted from

VHR optical satellite imagery. It must be noted that extraction at this level of detail for a large city like

Bamako is quite costly. Thematic detail (of nomenclature) can be reduced in favor of applicability of




semi-automated or fully automated approach. Discussion on such service cost balancing and trade-offs

is currently ongoing with WB team support as part of Phase 2 activities.

It is worth to note that further independent categories of urban green may be derived automatically:

distribution and density of green areas defined by presence of low or high vegetation (=land cover)

without distinguishing by its type (= land use).

The maps and graphs presented below have been designed to support spatial analysis of sustainable

public spaces network in Bamako by describing following characteristics:

Availability: share of public spaces out of a unit area

Accessibility: distance from a public space to nearest road

Connectivity: distance from a public space to the nearest neighboring public space

Inclusivity: share of population living within 400m distance

Calculation of following statistics was conducted on following subset of classes from OGA product

that best adhere to definition of green spaces with possible “public” function: Parks (downtown,

neighborhood), Squares (downtown, neighborhood) and Waterfronts.

Figure 15: Share of OGA: Total area of Open and Green Areas in wards (ha) – left; Share of Open and

Green Areas on Urban fabric per ward (%) – right, Bamako 2018

Figure 15 shows on the left the total area of urban green and open spaces (ha) per ward. Urban Green

areas used for this statistics were selected from Open Green classes. Figure on the right show Mean

ration of Urban Green and Open Spaces areas per Urban Fabric areas in wards.




Figure 16: Open Green Areas Connectivity: Mean distance between nearest Open Green Areas in wards

(m) – left; Euclidian distance to the nearest Open Green Areas calculated for each Open Green Space –

right; Bamako 2018

Figure 16 shows another insight into distribution of connectivity of the Open and Green Areas. The

figure on the left represents Mean distance between nearest Open Green Areas (OGA) per ward.

Figure on the right represents mutual distance relationships between Open Green Areas themselves -

each OGA is classified by Euclidian distance to its nearest OGA with indication of linear connection

links and their densities. This can support interpretation of fragmentation of selected OGA areas and

potential axes to support better connectivity – OGA infrastructure network within city.

Table 7: Inclusivity: Percentage of population living within 400m catchment areas recalculated for year

2018

Estimation of total

population in AOI (aggregated from

WorldPop 1000m

GRID)

Estimation of total

population within

400 m catchment

area

Share of estimated

population within 400m

catchment area on

estimated total per AOI

(%)

Bamako 2.559.737 990.439 38.69%




Figure 17: Open Green Areas Accessibility: Street path distance to the nearest Open Green Space (m) –

left; Average street path distance to the nearest Open Green Space (m) per ward – right; Bamako 2018

Figure 17 relates to Open and Green Areas accessibility. The image on the left shows distance to the

nearest OGA (via road network) – distance was calculated for each point in the network assuming all

roads and streets in the city while excluding highways. The figure on the right denotes average street

path distance to the nearest OGA aggregated from roads in each ward.




Spatial aspects of OGA distribution, now for full dataset without subsetting, within the city are further

analyzed using following maps and graphs. Relationship of selected variables on distances from the

city center - central business district (CBD) and directional axis defined by 30° transects is presented. CBD point was defined on one of central squares near the Bamako’s town hall (12°38’19” N, 7°59’48“ W).

Figure 18: Directional distribution and density (as % of Urban Fabric) of OGA classes within regular

distance bands the CBD in Bamako




Structure of OGA classes as a function of distance from CBD and ratio to urban fabric area is assessed

using following figures.

Figure 19: Structure of OGA as a function of distance from the CBD in Bamako




To better describe relationship of OGA’s size to the level of urbanization, which is not directly related to distance to CBD, Urbanity index has been introduced. The additive compound index consists of two

components calculated from spatial density of following factors within a 400m buffer zone around

each OGA polygon:

Normalized sealing values: range (0, maximum) rescaled into range (0, 1)

Normalized intersections density per sqkm: range (0, 100) rescaled into range (0,1), i.e. values

> 100 have automatically 1 as output.

The index values vary from the range (0, 2) with values closer to 2 indicating more urbanized areas.

Figure 20: Level of urbanity vs. public space size for two classes (cut-off at 99% percentile of public space

size)

4.5 Concluding Points

This Chapter 4 presented only a summary and overview of what is possible in term of analytics with

the geo-spatial datasets provided for Bamako in the current project. This Report is a living document

and will be complemented with further analysis during the project.




5 References

Czaplewski, R. L. (2003). Chapter 5: Accuracy assessment of maps of forest condition: statistical

design and methodological considerations, pp. 115–140. In Michael A. Wulder, & Steven E. Franklin

(Eds.), Remote sensing of forest environments: concepts and case studies. Boston: Kluwer Academic

Publishers (515 pp.).

European Union (2011). Mapping Guide for a European Urban Atlas, Version 11.0

Goodchild, M., Chih-Chang, L. and Leung, Y. (1994): Visualizing fuzzy maps, pp. 158-67. In

Heamshaw, H.H. and Unwin, D.J. (Eds.), Visualization in geographical information systems.

Chichester: Wiley.

Olofsson, P., Foody, G. M., Stehman, S. V., & Woodcock, C. E. (2013). Making better use of

accuracy data in land change studies: Estimating accuracy and area and quantifying uncertainty using

stratified estimation. Remote Sensing of Environment, 129, 122–131. doi:10.1016/j.rse.2012.10.031

Selkowitz, D. J., & Stehman, S. V. (2011). Thematic accuracy of the National Land Cover Database

(NLCD) 2001 land cover for Alaska. Remote Sensing of Environment, 115(6), 1401–1407.

doi:10.1016/j.rse.2011.01.020.

Internet

Road classification, European Commission, 2017, https://ec.europa.eu/transport/road_safety/specialist-

/knowledge/road/designing_for_road_function/road_classification_en, last accessed 2017.08.17

Earth Observation for Sustainable

Development Urban Project QA/QC Sheets developed by GAF AG

© 2018 GAF AG All Rights Reserved. Unless otherwise indicated, the templates of these QA/QC pages are copyrighted by GAF AG. No

part of these pages, either text or image may be used for any purpose other than use in the EO4SD-Urban Project. Therefore,

reproduction, modification, storage in a retrieval system or retransmission, in any form or by any means, electronic, mechanical or

otherwise, for reasons other than personal use, is strictly prohibited without prior written permission

Page 1








Page 2

Annex 1 – Processing Methods for EO4SD-Urban Products







Page 3








Page 4

Summary of Processing Methods

Urban Land Use/Land Cover

The input includes Very High Spatial Resolution (VHR) imagery from different sensors acquired at

different time. The data is pre-processed to ensure a high level of geometric and radiometric quality

(ortho-rectification, radiometric calibration, pan-sharpening).

The complexity when dealing with VHR images comes from the internal variability of the information

for a single land-use. For instance, an urban area is represented by a high number of heterogeneous

pixel values hampering the use of automated pixel-based classification techniques.

For these VHR images, it is possible to identify textures (or pattern) inside an entity such as an

agricultural parcel or an urban lot. In other words, whereas pixel-based techniques focus on the local

information of each single pixel (including intensity / DN value), texture analysis provides global

information in a group of neighbouring pixels (including distribution of a group intensity / DN values

but also spatial arrangement of these values). Texture and spectral information are combined with a

segmentation algorithm in an Object Based Image Analysis (OBIA) approach to reach a high degree

of automation for most of the peri-urban rural classes. However, within urban land, land use

information is often difficult to obtain from the imagery alone and ancillary/in situ data needs to be

used. The heterogeneity and format of these data mean that another information extraction method

based on Computer Aided Photo-Interpretation techniques (CAPI) need to be used to fully characterise

the LULC classes in urban areas. Therefore, a mix of automated (OBIA) and CAPI are used to

optimise the cost/quality ratio for the production of the LULC/LUCC product. The output format is

typically in vector form which makes it easier for integration in a GIS and for subsequent analysis.

Level 4 of the nomenclature can be obtained based on additional information. These can be generated

by more detailed CAPI (e.g. identification of waste sites) or by an automated approach based on

derived/additional products. An example is illustration by categorising the density of the urban fabric

which is related to population density and can then subsequently used for disaggregating population

data.

Information on urban fabric density can be obtained through several manners with increasing level of

complexity. The Imperviousness Degree (IMD) or Soil Sealing (SL) layer (see separate product) can

be produced relatively easily based on the urban extent derived from the LULC product and a linear

model between imperviousness areas and vegetation vigour that can be obtained from Sentinel 2 or

equivalent NDVI time series. This additional layer can be used to identify continuous and

discontinuous urban fabric classes. Five urban fabric classes can be extracted based on a fully

automated procedure:

Continuous Dense Urban Fabric (Sealing Layer-S.L. > 80%)

Discontinuous Dense Urban Fabric (S.L. 50% - 80%)

Discontinuous Medium Density Urban Fabric (S.L. 30% - 50%)

Discontinuous Low Density Urban Fabric (S.L. 10% - 30%)

Discontinuous Very Low Density Urban Fabric (S.L. < 10%)

Isolated Structures

Manual enhancement is the final post-processing step of the production framework. It will aim to

validate the detected classes and adjust classes’ polygon geometry if necessary to ensure that the correct MMU is applied. Finally, a thorough completeness and logical consistency check is applied to

ensure the topological integrity and coherence of the product.







Page 5


Settlement Extent

The rationale of the adopted methodology is that given a series of radar/optical satellite images for the

investigated AOI, the temporal dynamics of human settlements are sensibly different than those of all

other land-cover classes.

While addressing settlement-extent mapping for the period 2014-2015 we took into account

multitemporal S1 IW GRDH and Landsat-8 data acquired at 10 and 30m spatial resolution,

respectively. Concerning radar data, each S1 scene is pre-processed by means of the SNAP software

available from ESA; specifically, this task includes: orbit correction, thermal noise removal,

radiometric calibration, Range-Doppler terrain correction and conversion to dB values. Scenes

acquired with ascending and descending pass are processed separately due to the strong influence of

the viewing angle in the backscattering of built-up areas. As a means for characterizing the behaviour

over time, the backscattering temporal maximum, minimum, mean, standard deviation and mean slope

are derived for each pixel. Texture information is also extracted to ease the identification of lower-

density residential areas. As regards optical data, only Landsat-8 scenes with cloud cover lower than

60% are taken into consideration (indeed, further rising this threshold often results in accounting for

images with non-negligible misregistration error). Data are calibrated and atmospherically corrected

using the LEDAPS tool available from USGS and the CFMASK software is applied for removing

pixels affected by cloud-cover and cloud-shadow. Next, a series of 6 spectral indices suitable for an

effective delineation of settlements (identified through extensive experimental analysis) are extracted;

these include – among others – the Normalized Difference Built-Up Index (NDBI), the Modified

Normalized Difference Water Index (MNDWI) and the Normalized Difference Vegetation Index

(NDVI). For all of them, the same set of 5 key temporal statistics used in the case of S1 data are

generated for each pixel in the AOI. Moreover, to improve the detection of suburban areas, for each of

the 6 temporal mean indices also here texture information is computed. For matching the spatial

resolution of Sentinel data, the whole stack of Landsat-based features is finally resampled to 10m

spatial resolution.

To identify reliable training points for the settlement and non-settlement class, a strategy has been

designed which jointly exploits the temporal statistics computed for both S1 and Landsat data, along

with additional ancillary information. In the case of optical data, in general the most of settlement

pixels can be effectively outlined by properly jointly thresholding the corresponding NDBI, NDVI,

and MNDWI temporal mean; likewise, this holds also for non-settlement pixels. Regarding radar data,

it generally occurs that the temporal mean backscattering of most settlement samples is sensibly higher

than that of all other non-settlement classes. Nevertheless, in complex topography regions: i) radar

data show high backscattering comparable to that of urban areas; and ii) bare rocks are present, which

often exhibit a behaviour similar to that of settlements in the Landsat-based temporal statistics.

Accordingly, to exclude these from the analysis, all pixels are masked whose slope - computed based

on SRTM 30m DEM for latitudes between -60° and +60° and the ASTER DEM elsewhere - is higher

than 10 degrees.

Support Vector Machines (SVM) are used in the classification process. However, as the criteria

defined above for outlining training samples might results in a high number of candidate points, for

AOIs up to a size of ~10000 km² the most effective choice proved extracting 1000 samples for both the settlement and non-settlement class. Nonetheless, since results might vary depending on the

specific selected training points, as a means for further improving the final performances and obtain

more robust classification maps, 20 different training sets are randomly generated and given as input

to an ensemble of as many SVM classifiers; then, a majority voting is applied. Afterwards, the stacks

of Landsat-8-based and S1-based temporal features are classified separately as this proved more

effective than performing a single classification on their merger. In both cases, a grid search with a 5-

fold cross validation approach is employed to identify for each training set the optimal values for the







Page 6

learning. Here, those resulting in the highest cross-validation overall accuracy are then selected and

used for classifying the corresponding study region.

A final post-classification phase is dedicated to properly combining the Landsat- and S1-based

classification maps and automatically identifying and deleting potential false alarms. To this purpose,

an advanced post-editing object-based approach has been specifically designed.

The above-described methodology has been further adapted for outlining the settlement extent in the

past solely based on Landsat-5/7 imagery available since 1984; indeed, no long-term SAR data archive

at comparable spatial resolution is freely accessible for the same timeframe (e.g., ESA ERS-1/2 data

are available from 1991 without systematic world coverage and often proved too complicated to pre-

process). In particular, for the given target period and AOI, all available Landsat imagery with cloud

cover lower than 60% is pre-processed in the same fashion as described in the previous paragraphs and

the same set of temporal statistics and texture features are extracted. Based on the hypothesis that

settlement growth occurred over time (meaning that a pixel cannot be marked as settlement at an

earlier time if it has been defined as non-settlement at a later time), all pixels categorized as non-

settlement in the 2014-2015 extent map are excluded from the analysis. Then, training samples are

derived by thresholding the temporal mean NDBI, MNDWI and NDVI; specifically, a dedicated

strategy has been implemented for automatically determining the thresholds for the 3 indices by

comparing their cumulative distribution function (CDF) for the target period with that exhibited for the

period 2014-2015. Also in this case, an ensemble of 20 SVMs is used, each one trained on a different

subset of 2000 samples (i.e., 1000 for the settlement and 1000 for the non-settlement class) and

majority voting is then employed for generating the final map. It is worth noting that, when deriving

the past settlement extent for multiple times, both the masking and threshold adaptation are performed

on the basis of the results derived for the next target period.







Page 7


Transport Network

The transportation network is mainly manually digitised and mapped on the basis of very high

resolution (VHR) optical satellite imagery. Requested features are obtained by integration of auxiliary

data such as OpenStreetMap (OSM) or local datasets as a starting point for the product generation.

The revision and update of the auxiliary data is realised by using up-to-date VHR data. The workflow

can be specified as follows:

Quality check of available ancillary data such as local data sets.

Processing of optical satellite data – dependent on satellite data product level (geometric,

atmospheric and radiometric corrections, enhancements – colour optimization, mosaicking,

tiling).

Identification, collection and integration of available ancillary data (e.g. Open Street Map)

Identification and adjustment of spatial inconsistencies. The OSM data is used as spatial

reference. Upon User request other data sets can be used.

Update of the network by visual photo-interpretation according road hierarchy (see description

below).

Update of attributes by photointerpretation.

Generalization, application of MMU (minimum allowable dangling length)

Quality control and accuracy assessment

o Statistical sampling of check points

o Independent evaluation of products (second interpreter, third party assessment)

Change detection

The road hierarchy used in the classification is based on the international road classification

standards. One definition is specified by the European Commission. Roads are divided into three

groups - arterial or through traffic flow routes (in our case Arterial Roads), distributor road (in our

case Collector Roads), and access roads (in our case Local Roads). The three road types are

defined as follows: Arterial Roads - roads with a flow function allow efficient throughput of

(long distance) motorized traffic. All motorways and express roads as well as some urban ring

roads have a flow function. The number of access and exit points is limited. Collector Roads -

roads with an area distributor function allow entering and leaving residential areas, recreational

areas, industrial zones, and rural settlements with scattered destinations. Local Roads - roads with

an access function allow actual access to properties alongside a road or street. Arterial roads and

collector roads were the main focus of the classification. These types of roads were identified for

the entire Area of Interest.







Page 8


Urban Green Areas and Change

Generic methodology

Urban green areas refer to land within and on the edges of a city that is partly or completely covered

with grass, trees, shrubs, or other vegetation. This includes public parks, private gardens, cemeteries,

forested areas as well as trees, river alignments, hedges etc. The product delivered within EO4SD-

Urban project thus provides accurate information (1 m resolution) on the spatial location and extent of

the green areas located within the Urban Extent derived from the baseline LULC information product.

Detecting and monitoring urban green coverage needs VHR EO data, which explains the product

generation over the Core Urban Area of AOI only. The same images have been logically used as for

generating the LULC information product. Consequently, the usual preliminary quality check and pre-

processing tasks were already implemented.

Urban Green Areas have been detected using automated non-supervised classification method. More

precisely, each single multispectral VHR scene has been classified by specifying the most appropriate

algorithm and class number. Then, pixel units from the classes considered as representing green areas

have been combined into 1 single class. From this operation results the required binary raster product.

At this stage, it only remains necessary to apply some post-processing steps, such as eliminating the

pixel groups under the MMU, merging the products resulting from the classification process of each

single scene, but also integrating the information provided by the LULC product (class 14100 - Urban

Greenery). In case of reasonable data volume, the raster layer can be converted in vector format and

smoothing algorithm can be applied before delivery to users.

This processing method is applied for each reference date for which the product is required. Then,

change information layer is basically derived from the geometric intersection of the historic and

current Urban Green Areas layers by means of GIS operation. This resulting product finally provides

information about permanent vegetation, loss of urban green areas and new ones. Quality control and

accuracy assessment tasks are performed by means of visual interpretation and considering the LULC

dataset.

Specific methodology applied for OGA product

In Dhaka, extended typology for Urban Green areas was requested. While LULC at level 3 provides

only delineation of urban green, extended typology aims to stratify areas according to tailored

categorization addressing urban planning and management needs. Such typology includes also concept

of open spaces. Most of classes are derived within geometric outline of original LULC classes, but not

exclusively due to different application of land “use” concept in some cases (e.g. square may lay over multiple land cover / land use classes).

Nomenclature implementing the typology which is derivable from EO or ancillary data is detailed in

the following table.

Table 1-1: Open and Green Areas nomenclature

Level 1 Level 2

CODE_L1 DESC_L1 CODE_L2 DESC_L2 Remark

10 Square & open

11 Central city square Apparent squares or large crossings with visible

"public" function = shops, markets, sidewalks around

building fronts 12 Neighbourhood city square

13 Pocket city square

14 Suburban square







Page 9

15 Other open space Vacant open spaces within built-up fabric - land

without current visible (public or other) use. Open

spaces deemed to be associated land to other built-

up functional blocks. Open spaces in developing

zones at transitional urban fringe (subject to urban

sprawl) not included.

1 "Restricted" open Open spaces (without buildings) apparently enclosed

by wall or fence, probably publicly inaccessible

20 Market 21 Market (Open sky) Including large streets with concentration of stalls

30 Street 31 Street

Roads surrounded predominantly by urban

residential and non-residential fabric. Not included:

Capacity roads, Highways, Overpasses, Underpasses,

Road links

40 Waterfront

41 Beach Excluding mangrove forests, Industrial areas, Port

areas except berthing places for "public" river

transport 42 Waterfront park

43 Berthing & Boardwalk

50 Greenery

51 City park Horticultural management is directly visible from

imagery or very likely from land cover composition

52 Neighbourhood park

53 Pocket park

54 Linear park

55 Linear green Other elongated green

56 Trees, Sub-urban forest or

woodland Forest and trees within in city outline. No

horticultural management apparent

57 Other green Other unsorted green: trees, shrubs or grass

(residual class)

59 Airport green Greens within outline of airport area (associated

land).

60 Cemetery 61 Cemetery

70

Recreation, sport

and leisure

facilities

71 Recreational facilities Green sport areas incl. Buildings and associated land,

playgrounds, amusement parks, golf courses etc.

72 Leisure and amusement parks

73 Stadiums and sport facilities

74 Sport fields

90 Inaccessible open / green area

Methodology for deriving extended typology depends on utility of both VHR optical satellite data and

ancillary data from free sources – Open Street Map.

Green areas

Inputs: VHR Imagery (pan-sharpened multispectral), LULC product , Open Street Map (OSM)







Page 10

Green areas polygons from LULC product are taken as candidate polygons. Their geometry is revised

by interpretation of satellite image and assuming geometry of respective classes from OSM. Polygons

are further supplemented with small green area spot polygons which were originally derived by

automated approach but fell below the MMU threshold applicable for LULC product and now are are

above custom MMU applicable for green areas product.

Soil Sealing, Greenness, openness and size attributes are automatically calculated at a LULC block

level to supplement decision tree implemented in later stage.

Polygons are classified according to extended typology definition by expert interpretation. Information

visible in satellite imagery is considered as well as information from OSM. Public areas typically

cannot be distinguished from non-public with confidence as this information is not always obvious

from the imagery.







Page 11








Page 12

Annex 2 – Filled Quality Control Sheets







Page 13



Development – Urban Project QA/QC Sheets developed by GAF AG

© 2018 GAF AG All Rights Reserved. Unless otherwise indicated, the templates of these QA/QC pages are copyrighted by GAF AG. No part of these pages, either text or image may be used for any purpose other than use in the EO4SD-Urban Project. Therefore, reproduction, modification, storage in a retrieval system or retransmission, in any form or by any means, electronic, mechanical or otherwise, for reasons other than personal use, is strictly prohibited without prior written permission. Page 1

Earth Observation for Sustainable Development – Urban

Quality Assurance and Quality Control Sheets These QA/QC Templates were prepared by GAF AG compliant with ISO 9001:2015 Quality Management

System standards and can only be used by Partners in the current EO4SD-Urban Project.

Project Title: EO4SD-Urban

Project Leader: GAF AG

Service Provider: GISAT, S.R.O. Editor: Václav Stonáček

Client: WB Date: 21.2.2019

Product: LULC

Overview of QC-Sheets and Processing Steps Sheet used

Sheet filled in

Requirements

0.1 Requirements Yes Yes

Specifications of Input Data

1.1 List of EO Data Yes Yes

1.2 List of In-situ Data Yes No

1.3 List of Ancillary Data Yes Yes

Data Quality Checks

2.1 EO Data Quality Yes Yes

2.2 In-situ Data Quality Yes No

2.3 Ancillary Data Quality Yes Yes

Pre-Processing of EO Data

3.1 Geometric Correction Yes Yes

3.1.1 Data Fusion Yes Yes

3.2 Data Processing Yes Yes

Thematic Processing

4.1 Classification Yes Yes

4.2 Intermediate Quality Control of Land Use Data Yes Yes

Accuracy Assessment

5.1 Thematic Accuracy Yes Yes

5.2 Error Matrices Yes Yes

Delivery Checks / Delivery

6.1 Completeness Yes Yes

6.2 Compliancy Yes Yes




Glossary (index numbers in the QA/QC tables refer to the glossary at the end of this document) Further QC-relevant Documents:

Comments / Characteristics:




0.1 Requirements

Product 1 (28) Urban Land Use/ Land Cover

Abstract

The Land Use/Land Cover product contains spatial explicit information on different land use and land cover in Core Urban area of the city of Bamako (Mali) for the year of 2018. The Core area has detailed LU/LC nomenclature. The input data for the Core area was the Very High Resolution data (WorldView-3 for 2018). The LU/LC product is the Baseline Product from which various derived products (such as Open Green Areas) are produced.

Service / Product Specifications

Area Coverage

Country: Mali A) Wall-to-wall:

City: Bamako Selected Sites:

Area km² Core Urban: : 250 km2 B) Sampling based: n/a

Peri-Urban: N/A

Time Period - Update Frequency

A) Baseline Year(s): B) Update Frequency

2018 1 time horizon, no future update

Comments:

Geographic Reference System

UTM 29N

Mapping Classes and Definitions

Residential, 0 - 10 % Sealed

Urban fabric consist of individual houses with little evidence of apartment blocks. Plot size are relatively

large compare to formal high density residential. 90 % - 99% of the urban fabric consists of non-sealed or

vegetated surfaces.



large compare to formal high density residential. 70 % - 90 % of the urban fabric consists of non-sealed or

vegetated surfaces.



large compare to formal high density residential. 50 % - 70 % of the urban fabric consists of non-sealed or

vegetated surfaces.


Urban fabric where formal housing (in the form of individual houses or apartment blocks) dominates. Formal

houses expected to be organized in a relatively regular spatial pattern with clearly visible roads. 20 %-50 % of

the urban fabric consists of non-sealed or vegetated surface.


Urban fabric where formal housing (in the form of individual houses or apartment blocks) dominates. Formal

houses expected to be organized in a relatively regular spatial pattern with clearly visible roads. 0 %-20 % of

the urban fabric consists of non-sealed or vegetated surface.

Commercial Area Shopping malls, markets and all associated facilities and land includes medium to large scale compound

units.




Industrial Area Factories, warehouses, kilns and all associated facilities and land includes medium to large scale compound

units.

University University area, with larger buildings, often with high vegetation.

Schools Schools buildings with playground or other associated areas.

Government Governmental buildings.

Military Military areas or buildings, mostly fenced.

Hospitals Larger buildings serve for healing people, usually with high vegetation, sometimes with heliport.

Public Buildings Buildings serve for public use, as libraries, churches, banks, etc.

Non-Residential Urban Fabric

Buildings that do not serve for residential, but without any information about use.

Arterial Line Motorways and main city roads and associated land.

Collector Line Primary and secondary roads and associated land.

Railway Railway and associated land.

Airport Airport and associated land

Port Port and associated land.

Mining Area and Dump Sites

Areas used for mining minerals or dump sites.

Construction site Areas with on-going building/infrastructure construction activity or areas obviously prepared for

construction.

Vacant Land not obviously being prepared for construction

Vacant land for which there is no evidence of on-going building/infrastructure construction activity.

Urban Parks (Urban Green Spaces)

Urban green area.

Recreation Facilities Recreational use as gardens, zoos, parks including sport and leisure facilities, sport fields.

Cemeteries Cemeteries

Agricultural Area Cultivated areas non-irrigated or permanently irrigated including rice fields. Arable land (annual crops)

Forest and Shrub Lands High woody vegetation in natural forests including bushes and shrubs at the fringe of the forest.

Natural Areas Herbaceous vegetation, grassland.

Bare Soil Natural areas where there is no or very little evidence of vegetation and does not serve as construction site.

Wetlands Wetlands, mangrove and other water plants.

Water Visible water surface areas.

Cloud and Cloud Shadow Detection and Removal

EO data without clouds or shadows.

Spatial Resolution

n.a. (Product provided as Shapefile)

Minimum Mapping Unit (MMU)

Minimum Mapping Unit is 0.25 ha for Core City

Data Type & Format

Vector files (Shapefile), Maps (JPEG, TIFF, PDF)

Bit Depth

N/A




Class Coding

Class Code Class Name RGB Code

1100 Residential 255; 0; 0 (main class only)

1110 Residential, 80 - 100 % Sealed 115; 0; 0





1211 Commercial Area 197; 0; 255

1212 Industrial Area 132; 0; 168

1213 University 194; 158; 215

1214 Schools 232; 190; 255

1215 Government 115; 223; 255

1216 Military 0; 255; 197

1217 Hospitals 115; 115; 0

1218 Public Buildings 192; 252; 234

1219 Non-Residential Urban Fabric 168; 0; 132

1221 Arterial Line 78; 78; 78

1222 Collector Line 120; 120; 120

1223 Railway 52; 52; 52

1230 Port 0; 168; 132

1240 Airport 79; 0; 115

1310 Mining Area and Dump Sites 115; 76; 0

1330 Construction site 255; 115; 223

1340 Vacant Land not obviously being prepared for construction 242; 242; 242

1410 Urban Parks (Urban Green Spaces) 85; 255; 0

1420 Recreation Facilities 255; 170; 0

1430 Cemeteries 230; 230; 0

2000 Agricultural Area 255; 235; 175

3100 Forest and Shrub Lands 38; 115; 0

3200 Natural Areas 180; 215; 158

3300 Bare Soil 204; 204; 204

4000 Wetlands 76; 0; 115

5000 Water 0; 112; 255

Metadata

Provided as INSPIRE conformant *.xml data set, covering at least the mandatory elements.

Service / Product Quality

Thematic Accuracy

Overall Accuracy: 93%

Positional Accuracy

RMSE < 30 m.




Delivery Procedure

Service Provision

Online via FTP

Delivery Date

March 2019.




1.1 List of EO Data

Sensors (8) WorldView-3

Incoming Date

Acquisition Date

Proc. Level

Path / Row

AOI - City / Region / Country

Spatial Res.

No. of Bands

Cloud Cover (Data Provider)

Projection / Spheroid (16)

Data Format (3)

Bit Depth (5)

Header / Metadata (2)

File Name [e.g yymmdd; tbd...]

WorldView-3

1. 058525860010_01_P001 15.10.2018

4.1.2018

2A N/A Bamako MUL:2 m PAN: 0.5 m

4 0% UTM29 / WGS84

*.tif 16 Bit u

*.imd / *.xml

2. 058525860010_01_P002 15.10.2018

14.1.2018


4 0% UTM29 / WGS84

*.tif 16 Bit u

*.imd / *.xml

3. 058525860010_01_P003 15.10.2018

14.1.2018


4 0% UTM29 / WGS84

*.tif 16 Bit u

*.imd / *.xml




1.2 List of In-situ Data

N/A

1.3 List of Ancillary Data

Dataset 1 (14)

DEM Date of Receipt from Client

Metadata (2)

Reference Mapping Date / Date of Creation

Geographic Area/- City / Region / Country

Area Coverage

Data Type (4)

Data Format (3)


Positional Accuracy (11)

No. of Classes

Thematic Accuracy

Availability of Class Definitions

SRTM1arcsec 9.10.2018 *.mta 11.-22.02. 2000

Bamako/Mali 100% Raster *.tif WGS84

Horizontal: CE90 <20m Vertical: CE90 <16m

N/A unknown N/A

Lineage(29):

The Shuttle Radar Topography Mission (SRTM) was flown aboard the space shuttle Endeavour on February 11 - 22, 2000. The National Aeronautics and Space Administration (NASA) and the National Geospatial-Intelligence Agency (NGA) participated in an international project to acquire radar data which were used to create detailed topographic maps. SRTM data are intended for scientific use with a Geographic Information System (GIS) or other special application software. Endeavour orbited Earth 16 times each day during the 11-day mission completing 176 orbits. SRTM successfully collected radar data over 80% of the Earth's land surface between 60° north and 56° south latitude with data points posted every 1 arc-second (approximately 30 meters).

Source (30):: http://dds.cr.usgs.gov/srtm/version2_1/SRTM3

Dataset 2 (14)

OSM Date of Receipt from Client

Metadata (2)


Geographic Area / City / Region / Country

Area Coverage

Data Type (4)

Data Format (3)

Projection / Spheroid

(16)


No. of Classes

Thematic Accuracy

Class Definitions

OSM roads/railways/waterways

8.8.2018 N/A N/A Bamako/Mali 100% Vector *.shp UTM29 / WGS84

N/A N/A unknown Yes

Lineage(29): Up-to-date data extracts from the OpenStreetMap project. OSM data is crowded source data with contribution from volunteers and freely available.

Source (30):: Htttp://www.download.geofabrik.de

http://dds.cr.usgs.gov/srtm/version2_1/SRTM3




2.1 EO Data Quality

Sensors(8) WorldView-3

Ba

cku

p

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/

Me

tad

ata

(2)

Band Specifications

Pro

ject

ion

/ S

ph

ero

id (1

6)

Sce

ne

Lo

cati

on

Co

mp

lete

ne

ss o

f A

dd

itio

na

l D

ata

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Clo

ud

Co

ver

(in

tern

al

che

ck)

Ra

dio

me

try

To

po

gra

ph

y

Dro

pp

ed

Lin

es

/ A

rte

fact

s (7

)

Acc

ep

tan

ce S

tatu

s

Comment

s File Name [e.g yymmdd; tbd...]

No

. o

f B

an

ds

Ba

nd

Re

gis

tra

tio

n

Sp

ect

ral/

Sp

ati

al

Re

solu

tio

n

WorldView-3

1. 058525860010_01_P001 Y - Q

Y - V Q

Y - V Q

Y - V Q

Y - V Q

Y - Q

Y - V Y - V Y - V

Q Y - Q

Y - Q

Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

2. 058525860010_01_P002 Y - Q

Y - V Q

Y - V Q

Y - V Q

Y - V Q

Y - Q

Y - V Y - V Y - V

Q Y - Q

Y - Q

Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

3. 058525860010_01_P003 Y - Q

Y - V Q

Y - V Q

Y - V Q

Y - V Q

Y - Q

Y - V Y - V Y - V

Q Y - Q

Y - Q

Y - V Q

Y - V Q

Y - V

N - V Q

Yes none




2.2 In-situ Data Quality

N/A




2.3 Ancillary Data Quality

Dataset 1 (14) DEM

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

M

atc

he

s S

erv

ice

Are

a (

%)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

a

nd

un

it (

e.g

. m

, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/ A

rte

fact

s (

EO

da

ta

on

ly)(

7)

Uti

lity

for

Cu

rre

nt

Pro

ject


SRTM1arcsec ☒ Yes

☐ No

☐ Complete (INSPIRE/ISO19119)

☐ Incomplete

☒ not available

☒ Yes

☐ No

☐ Unknown If no: xxx%

☒ Correct

☐ Incorrect

☐ Unknown EPSG: 4326

30 m *.tiff 16 bit S

☒ Complete

☐Incomplete

☒ No

☐ Yes

☐ Unknown If yes: XX m

☒ No

☐ Yes

☐ n.a. If yes: xxx %

☐ None

☐ Partial

☒ Full

Comments: Due to missing Metadata the Accuracy and Completeness of the data set is unknown.

Dataset 2 (14) OSM

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

Ma

tch

es

Se

rvic

e

Are

a (

%)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

an

d u

nit

(e

.g.

m, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/ A

rte

fact

s (

EO

da

ta

on

ly)(

7)

Uti

lity

for

Cu

rre

nt

Pro

ject


OSM roads/railways/waterways ☒ Yes

☐ No


☐ Incomplete

☒ not available

☒ Yes

☐ No


☒ Correct

☐ Incorrect


n.a. *.shp n.a.

☒ Complete

☐Incomplete

☒ No

☐ Yes


☒ No

☐ Yes


☐ None

☐ Partial

☒ Full

Comments:

3.1 Geometric Correction




Sensor (8) WorldView-3

Pro

cess

ing D

ate

AO

I

Cit

y /

Re

gio

n /

Co

un

try

Pro

ject

ion

/ S

ph

ero

id (1

6)

No

. &

RM

S (

m)

of

GC

Ps

(17

)

No

. &

RM

S (

m)

of

TP

s (1

8)

No

. &

RM

S (

m)

of

CP

s (1

9)

Dig

ita

l E

leva

tio

n M

od

el (D

EM

)

Mo

de

l /

Alg

ori

thm

(13

)

Re

sam

plin

g

Me

tho

d (2

0)

Va

lida

tio

n R

ep

ort

s

Acc

ep

tan

ce S

tatu

s

Output File Name [e.g yymmdd; tbd...]

No. File Name [e.g yymmdd; tbd...]

WorldView-3

1. 058525860010_01_P001 17.10.2018

Bamako

UTM29 / WGS84

9; X RMSE 0.66, Y RMSE 0.62

N/A N/A SRTM

30

polynomic; 1st order

CUB

Yes Yes o058525860010_01_

P001

2. 058525860010_01_P002 17.10.2018

Bamako

UTM29 / WGS84

9; X RMSE 0.78, Y RMSE 0.37

N/A N/A SRTM

30


CUB

Yes Yes o058525860010_01_

P001

3. 058525860010_01_P003 17.10.2018

Bamako

UTM29 / WGS84

10; X RMSE 0.56, Y RMSE 0.46

N/A N/A SRTM

30


CUB

Yes Yes o058525860010_01_

P001




3.1.1 Data Fusion

Dataset 1 (14) Input A Input B Output

Acc

ep

tan

ce S

tatu

s

Comments

Filename 058525860010_01_P001_M

UL 058525860010_01_P001_PA

N 058525860010_01_P001_PS

Sensor WorldView-3 WorldView-3 WorldView-3

Method PANSHARP2 (PCI) PANSHARP2 (PCI) PANSHARP2 (PCI)

Spatial Resolution / (MS/Pan)

2 m MS 0,5 m Pan 0,5 m PS

Band Combination R,G,B,Nir PAN R,G,B,Nir

Data Format (3) &

Bit Depth (5) *.img 16 Bit u *.img *.16 Bit u *.tiff

16 Bit u

Yes


Acc

ep

tan

ce S

tatu

s

Comments

Filename 058525860010_01_P002_M

UL 058525860010_01_P002_PA

N 058525860010_01_P002_PS






Data Format (3) &


16 Bit u

Yes


Acc

ep

tan

ce S

tatu

s

Comments

Filename 058525860010_01_P003_M

UL 058525860010_01_P003_PA

N 058525860010_01_P003_PS






Data Format (3) &


16 Bit u

Yes




3.2 Data Processing

Sensor (8) WorldView-3 Atmospheric Correction

Radiometric Processing (15)

Topographic Normalisation

OTHER Calculation:

Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment

No. Processing

Date

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

WorldView-3

1. o058525860010_01_P001 17.10.2018 N N/A N N/A N N/A N N/A Yes Yes none






4.1 Classification


Processing

Date

Cloud Masking

Thematic Classification

Manual Enhancement

Border Matching

Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment N

o. File Name [e.g yymmdd; tbd...]

pro

cess

ed

Software /

Method p

roce

sse

d

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software /

Method

WorldView-3

1. o058525860010_01_P001 30.10.2

018 No N/A Yes

eCognition/semi-automatic approach

Yes

Webim/GISAT proprietary solution

No N/A Yes

Yes

2. o058525860010_01_P001 30.10.2

018 No N/A Yes


Yes


No N/A Yes

Yes

3. o058525860010_01_P001 30.10.2

018 No N/A Yes


Yes


No N/A Yes

Yes




5.1 Thematic Accuracy

No. Product Name

Processing Date

Area Coverage

No

. o

f C

lass

es

Re

fere

nce

Da

ta

Sa

mp

le S

ize

Sa

mp

ling U

nit

(21

)

Sa

mp

ling

De

sign

(22

)

Sa

mp

le E

xclu

sio

n

Cri

teri

a (2

3)

Th

em

ati

c A

ccu

racy

(2

4)

Acc

ep

tan

ce S

tatu

s

Comment

1.

Land Cover Map 2018

22.2.2019 Bamako/Mali (250km²)

9 VHR image 2018

460 POINT / pixel centre

STRAT N/A 93% Yes




5.2 Error Matrices

Class No Land Use Map 2018

Totals

1 2 3 4 5 6 7 8 9

Reference Data

Residential Urban

Fabric

Industrial, Commercial,

Public, Military, Private and

Transport Units

Roads and rail network

and associated land

Mine, Dump and

Construction Sites

Artificial non-agricultural vegetated

areas

Agricultural Area

Natural and

Semi-natural Areas

Wetlands

Water

1 Residential Urban Fabric 173 1 0 3 0 0 0 0 0 177

2 Industrial, Commercial, Public, Military, Private and Transport Units

7 82 0 1 1 0 0 0 1 92

3 Roads and rail network and associated land 0 1 10 0 0 0 0 0 0 11

4 Mine, Dump and Construction Sites 3 1 1 43 0 0 3 0 1 52

5 Artificial non-agricultural vegetated areas 0 0 0 1 24 1 0 0 0 26

6 Agricultural Area 1 0 0 0 0 36 0 0 0 37

7 Natural and Semi-natural Areas 1 0 0 1 0 3 45 0 0 50

8 Wetlands 0 0 0 0 0 0 0 4 0 4

9 Water 0 0 0 0 0 0 0 0 11 11

Totals 185 85 11 49 25 40 48 4 13 460




Accuracy Statistics z= 1,96


95% Confide

nce Interval:

90,7% 95,4%

Kappa (Cohen): 0,85

Class Name Producer Accuracy

95% Confidence Interval User

Accuracy 95% Confidence

Interval

Residential Urban Fabric 97,7% 96,4% 99,1% 93,5% 91,3% 95,8%

Industrial, Commercial, Public, Military, Private and Transport Units

89?1% 86,3% 92,0% 96,5% 94,8% 98,2%

Roads and rail network and associated land 90,9% 88,3% 93,5% 90,9% 88,3% 93,5%

Mine, Dump and Construction Sites 82,7% 79,2% 86,1% 87,8% 84,8% 90,8%

Artificial non-agricultural vegetated areas 92,3% 89,9% 94,7% 96,0% 94,2% 97,8%

Agricultural Area 97,3% 95,8% 98,8% 90,0% 87,3% 82,7%

Natural and Semi-natural Areas 90,0%% 87,3% 92,7% 93,8% 91,5% 96,0%

Wetlands 100,0% 100,0% 100,0% 100,0% 100,0% 100,0

%

Water 100,0% 100,0% 100,0% 84,6% 81,3% 87,9%




6.1 Completeness

INPUT DATA

No.

Item AOI Coverage [km²]

Area Coverage [km²]

Completeness of Coverage (25)

No. of Scenes Scenes used in Production

Completeness of Verification (27)

Metadata Comments

1. VHR EO Data Bamako (250 km2)

250km² 100% 3 3 100% Yes none

2. Ancillary Data N/A N/A N/A N/A N/A Yes none

PRODUCTS

No.

Product (28) AOI Coverage [km²]

Product Coverage [km²]


Completeness of Classification (26)

Unclassifiable Area [%]


Metadata Comments

1. LULC 2018 Bamako (250

km2) 250 km² 100% 100% N/A 100% Yes none




6.2 Compliancy

Product 1 (28) Urban Land Use/ Land Cover

Abstract

The Land Use/Land Cover product contains spatial explicit information on different land use and land cover in Core Urban area of the city of Bamako (Mali) for the year of 2018. The Core area has detailed LU/LC nomenclature. The input data for the Core area was the Very High Resolution data (WorldView-3 for 2018). The LU/LC product is the Baseline Product from which various derived products (such as Open Green Areas) are produced.


Area Coverage

Requirements Achieved Specifications Compliancy Comments

Country 1: Mali Country 1: Mali Yes

Country 2: N/A Country 2: N/A ---

A) Wall-to-wall: A) Wall-to-wall: Yes

Sub-national, Bamako Sub-national, Bamako

Area of approx. 250 km², defined by the national user.

B) Sampling based: B) Sampling based:

N/A N/A



Baseline Year(s): Baseline Year(s): Yes

Most actual 2018

B) Update Frequency B) Update Frequency

N/A N/A



WGS84 / UTM 29N WGS84 / UTM 29N Yes ---

Mapping Classes and Definitions (Definitions see REDD+ MRV Design Document)

Requirements Achieved Specifications Compliancy

31 Main classes (see 0.1 Requirements sheet) All classes were mapped according to their definition.

yes

Cloud and Shadow Detection and Removal


Cloud free. Cloud free scenes used. Yes

Spatial Resolution





VHR for core VHR for core Yes



0.25 ha for Core City 0.25 ha for Core City Yes

Data Type


Vector data ESRI Shapefile Yes ---

Bit Depth


Not specified N/A N/A ---

Data Format


*.shp Shapefile ESRI Shapefile Yes ---

(open cross-platform format)

Class Coding (Raster Data Only)


Na Na ---

Metadata


Metadata ISO compliant INSPIRE compliant and attached to

product Yes ---


Thematic Accuracy


Overall Accuracy: 90% - 95% (85% +/- 5% according to DNV recommendation)

93% Yes

Positional Accuracy


RMSE < 30 m RMSE < 2 m

Delivery Procedure

Service Provision


online via FTP or DVD, Harddisk FTP Yes

Delivery Date





March 2019




Glossary

Quality Checks Prefix Suffix Explanation

Y (Yes) - V Visually checked

N (No) - Q Quantitatively/qualitatively checked

N/A - V Q Visually and quantitatively/qualitatively checked

- N/A Not applicable

Readability (1) Check readability of all required input data. Can the data be stored again?

Header / Metadata (2) Check Image Header Information and/or Metadata for completeness / distinctive features.

Data Format (3) For digital data, please give file format (e.g. *.tiff, *.shp).

Data Type (4) Please specify the type of the data (e.g. raster, vector or analogue).

Bit Depth (5) Please give pixel depth and sign of raster data (e.g. 8 bit unsigned integer).

Dynamic Range (6) Check dynamic range of all image bands. Visual check of dynamic range should be accompanied by histograms and statistics.

Dropped Lines & Artefacts (7) Check Image for dropped lines and other artefacts. If such occurs, please give description of extent and influence in the Comments section.

Sensor 1, 2, … (8) Please delete / add additional sensor sections as necessary.

Purpose (9)

The purpose and use of the given auxiliary or reference data should be stated:

ORHTOrectification, GEOmetric correction, REFerence, VERification source, POSitional ACCuracy assessment, THEmatic ACCuracy assessment.

Sampling Methodology (10)

Methodology of in situ or reference/auxiliary data sampling scheme should be outlined. In case of sample plots, also state how the plot positions have been determined (e.g. from GPS measurements, topographic maps, terrestrial triangulation, EO data, etc.) and how the sampling grid was established.


Positional accuracy of collected in situ data should be given. For reference/auxiliary data it MUST be given. If unknown, the data’s use must be explained. For DEM or other data with 3D information please specify both vertical and horizontal Positional Accuracy. For analogue data (e.g. maps) try to give approximate accuracy related to mapping scale.

Completeness (12) Data should be checked for spatial/temporal/content gaps.

Model / Algorithm (13)

Give the name of the software and its version. Specify the software module/algorithm used for: a) geometric correction, e.g., Polynomial and its degree, Rational Functions, Thin Plate Spline, etc. b) classification, e.g., ISODATA, Maximum Likelihood, Neural Networks, etc.

Dataset 1, 2, … (14) Please delete / add additional dataset sections as necessary.





State whether (and which) radiometric processing was applied (e.g., contrast enhancement, histogram matching, bundle block adjustment, radiometric normalisation, filtering, etc.).

Projection; Spheroid / Ellipsoid (16)

Always specify completely, i.e. at minimum the Projection (+Zone, if applicable), Spheroid / Ellipsoid, Map Datum. Give additional information if necessary to unambiguously define the reference system.

Ground Control Points (GCPs) (17)

Give the number, distribution and RMS of used Ground Control Points (as average per scene) as obtained from the Geometric Correction, in meters [m]. Optionally, also RMSx and RMSy may be given. Distribution of GCPs should be attached as a snapshot, or described in the Comments section.

Tie Points (TPs) (18)

In case of mosaicking, give the number of used Tie Points and their total RMS (as average per scene) as obtained from the Geometric Correction, in meters [m]. Optionally, also RMSx and RMSy may be given. Distribution of TPs should be attached as a snapshot, or described in the Comments section.

Check Points (CPs) (19)

The real measure of positional accuracy and the only measure which should be examined as to its Acceptable Range. Give the independent Check Points’ total RMS (as average per scene) as obtained from the Geometric Correction, in meters [m]. Optionally, also RMSx and RMSy may be given. Distribution of CP's should be attached as a snapshot, or described in the Comments section.

Resampling Method (20) Specify, if / which resampling algorithm has been used (e.g. NN=Nearest Neighbour, BIL=Bilinear Interpolation, CC=Cubic Convolution).

Sampling unit (21) Specify sampling unit of Accuracy Assessment as POINT, FRAME, POLYGON and how it is treated (e.g. pixel center, polygon centre, etc.).

Sampling Design (22) SYST=Systematic, RAND=Random, STRAT=Stratified, SBCLASS=Stratified by class, SBAREA=Stratified by area

Sample exclusion criteria (23)

Describe which criteria you apply for sampling point selection resp. exclusion of certain points. For example, if the point is too close to a class boundary (less than 1 pixel), it is excluded. If the selected sample point is not representative of the class , it is excluded. For these reasons, it is recommended to oversample by 10% to compensate for sample point exclusion.

Thematic Accuracy (24)

Provide a detailed description of the Accuracy Assessment results in the form of error matrices showing commission and omission errors, user’s , producer’s and overall accuracies and other measures of Thematic Accuracy, as the confidence level (usually fixed at 95%) and the respective confidence interval, at least for the overall accuracy. If classification is done in a phased approach, e.g. if Forest Area and subsequently Forest Type are mapped, independent reports have to be produced.

Completeness of Coverage (25) State whether the coverage is limited to a subset, or portion of the final product.

State whether classification was constrained to a subset, or portion of the final product.






State whether verification applied to lineage, positional, or thematic accuracy is constrained to a subset, or portion of the final product.

Product (28) Please delete / add additional product sections as necessary.

Lineage (29)

Definition (INSPIRE 2015): Lineage is “a statement on process history and/or overall quality of the spatial data set. Where appropriate it may include a statement whether the data set has been validated or quality assured, whether it is the official version (if multiple versions exist), and whether it has legal validity. The value domain of this element is free text.”

Source(30) Definition (INSPIRE, 2015): “This is the description of the organisation responsible for the establishment, management, maintenance or distribution of the resource. This description shall include: name of the organisation and contact email address.”











Product: LULC


Sheet filled in

Requirements






Data Quality Checks








Thematic Processing



Accuracy Assessment














0.1 Requirements

Product 1 (28) Urban Green Areas 2018

Abstract

Urban green areas refer to land within and on the edges of a city that is partly or completely covered with grass, trees, shrubs, or other vegetation. This includes public parks, private gardens, cemeteries, forested areas as well as trees, river alignments, hedges, etc. The product delivered within EO4SD-Urban project thus provides accurate information (1m resolution) on the spatial location and extent of the green areas located within the Urban Extent derived from baseline LULC information product.


Area Coverage




Peri-Urban: N/A




Comments:


UTM 29N


Square & Open

Central city square

Apparent squares or large crossings with visible "public" function = shops, markets, sidewalks

Neighborhood city square

Pocket city square

Suburban square

Other open spaces Vacant open spaces within built-up fabric - land

without current visible (public or other) use.

Market Market (Open sky) Including large strees with concentration of stalls

Waterfront Excluding mangrove forests, Industrial areas, Port

areas except berthing places for "public" river

Greenery

City park

Horticultural management is directly visible from imagery or very likely from land cover

Neighborhood park

Pocket park

Linear park




Linear green Other elongated green

Tree, Sub-urban forest or woodland Forest and trees within in city outline. No

horticulturel management apparent

Other green Other unsorted green: trees, shrubs or grass

(residual class)

Cemetery

Recreation, sport and leisure facilities

Green sport areas incl. Buildings and associated

land, playgrounds, amusement parks, golf

Inaccessible open/green area



Spatial Resolution




Data Type & Format


Bit Depth

N/A

Class Coding


11 Square - City

255;0;0

12 Square- Neighborhood 255;85;0

13 Square - Pocket 255;167;127

14 Square - suburban 255;190;190

15 Other open space 255;255;115

21 Market (Open Sky) 135;38;117

40 Waterfront 115;178;255

51 Park - City 56;168;0

52 Park - Neighborhood 76;230;0

53 Park - Pocket 209;255;115

54,55 Green - Linear 0;255;197

56 Green - Trees, Forest, Woodland 38;115;0

57 Green - Other 233;255;190

61 Cemetery 232;190;255

70 Green - Sport & Leisure 168;112;0

90 Inaccessible potential PS 0;0;0

Metadata



Thematic Accuracy

Overall Accuracy: 94,3%




Positional Accuracy

RMSE < 30 m.

Delivery Procedure

Service Provision

Online via FTP

Delivery Date

March 2019.




1.1 List of EO Data


Incoming Date

Acquisition Date

Proc. Level

Path / Row


Spatial Res.

No. of Bands



Data Format (3)

Bit Depth (5)



WorldView-3

1. 058525860010_01_P001 15.10.2018

4.1.2018


4 0% UTM29 / WGS84

*.tif 16 Bit u

*.imd / *.xml

2. 058525860010_01_P002 15.10.2018

14.1.2018


4 0% UTM29 / WGS84

*.tif 16 Bit u

*.imd / *.xml

3. 058525860010_01_P003 15.10.2018

14.1.2018


4 0% UTM29 / WGS84

*.tif 16 Bit u

*.imd / *.xml





N/A


Dataset 1 (14)


Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy


SRTM1arcsec 9.10.2018 *.mta 11.-22.02. 2000



N/A unknown N/A

Lineage(29):



Dataset 2 (14)


Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)


(16)


No. of Classes

Thematic Accuracy

Class Definitions



N/A N/A unknown Yes







2.1 EO Data Quality


Ba

cku

p

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/

Me

tad

ata

(2)

Band Specifications

Pro

ject

ion

/ S

ph

ero

id (1

6)

Sce

ne

Lo

cati

on

Co

mp

lete

ne

ss o

f A

dd

itio

na

l D

ata

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Clo

ud

Co

ver

(in

tern

al

che

ck)

Ra

dio

me

try

To

po

gra

ph

y

Dro

pp

ed

Lin

es

/ A

rte

fact

s (7

)

Acc

ep

tan

ce S

tatu

s

Comment


No

. o

f B

an

ds

Ba

nd

Re

gis

tra

tio

n

Sp

ect

ral/

Sp

ati

al

Re

solu

tio

n

WorldView-3

1. 058525860010_01_P001 Y - Q

Y - V Q

Y - V Q

Y - V Q

Y - V Q

Y - Q

Y - V Y - V Y - V

Q Y - Q

Y - Q

Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

2. 058525860010_01_P002 Y - Q

Y - V Q

Y - V Q

Y - V Q

Y - V Q

Y - Q

Y - V Y - V Y - V

Q Y - Q

Y - Q

Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

3. 058525860010_01_P003 Y - Q

Y - V Q

Y - V Q

Y - V Q

Y - V Q

Y - Q

Y - V Y - V Y - V

Q Y - Q

Y - Q

Y - V Q

Y - V Q

Y - V

N - V Q

Yes none





N/A





Dataset 1 (14) DEM

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

M

atc

he

s S

erv

ice

Are

a (

%)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

a

nd

un

it (

e.g

. m

, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/ A

rte

fact

s (

EO

da

ta

on

ly)(

7)

Uti

lity

for

Cu

rre

nt

Pro

ject


SRTM1arcsec ☒ Yes

☐ No


☐ Incomplete

☒ not available

☒ Yes

☐ No


☒ Correct

☐ Incorrect



☒ Complete

☐Incomplete

☒ No

☐ Yes


☒ No

☐ Yes


☐ None

☐ Partial

☒ Full


Dataset 2 (14) OSM

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

Ma

tch

es

Se

rvic

e

Are

a (

%)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

an

d u

nit

(e

.g.

m, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/ A

rte

fact

s (

EO

da

ta

on

ly)(

7)

Uti

lity

for

Cu

rre

nt

Pro

ject



☐ No


☐ Incomplete

☒ not available

☒ Yes

☐ No


☒ Correct

☐ Incorrect


n.a. *.shp n.a.

☒ Complete

☐Incomplete

☒ No

☐ Yes


☒ No

☐ Yes


☐ None

☐ Partial

☒ Full

Comments:






Pro

cess

ing D

ate

AO

I

Cit

y /

Re

gio

n /

Co

un

try

Pro

ject

ion

/ S

ph

ero

id (1

6)

No

. &

RM

S (

m)

of

GC

Ps

(17

)

No

. &

RM

S (

m)

of

TP

s (1

8)

No

. &

RM

S (

m)

of

CP

s (1

9)

Dig

ita

l E

leva

tio

n M

od

el (D

EM

)

Mo

de

l /

Alg

ori

thm

(13

)

Re

sam

plin

g

Me

tho

d (2

0)

Va

lida

tio

n R

ep

ort

s

Acc

ep

tan

ce S

tatu

s



WorldView-3

1. 058525860010_01_P001 17.10.2018

Bamako

UTM29 / WGS84

9; X RMSE 0.66, Y RMSE 0.62

N/A N/A SRTM

30


CUB

Yes Yes o058525860010_01_

P001

2. 058525860010_01_P002 17.10.2018

Bamako

UTM29 / WGS84

9; X RMSE 0.78, Y RMSE 0.37

N/A N/A SRTM

30


CUB

Yes Yes o058525860010_01_

P001

3. 058525860010_01_P003 17.10.2018

Bamako

UTM29 / WGS84

10; X RMSE 0.56, Y RMSE 0.46

N/A N/A SRTM

30


CUB

Yes Yes o058525860010_01_

P001




3.1.1 Data Fusion


Acc

ep

tan

ce S

tatu

s

Comments

Filename 058525860010_01_P001_M

UL 058525860010_01_P001_PA

N 058525860010_01_P001_PS






Data Format (3) &


16 Bit u

Yes


Acc

ep

tan

ce S

tatu

s

Comments

Filename 058525860010_01_P002_M

UL 058525860010_01_P002_PA

N 058525860010_01_P002_PS






Data Format (3) &


16 Bit u

Yes


Acc

ep

tan

ce S

tatu

s

Comments

Filename 058525860010_01_P003_M

UL 058525860010_01_P003_PA

N 058525860010_01_P003_PS






Data Format (3) &


16 Bit u

Yes




3.2 Data Processing




OTHER Calculation:

Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment

No. Processing

Date

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

WorldView-3







4.1 Classification


Processing

Date

Cloud Masking


Manual Enhancement

Border Matching

Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment N


pro

cess

ed

Software /

Method p

roce

sse

d

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software /

Method

WorldView-3

1. o058525860010_01_P001 30.10.2

018 No N/A Yes


Yes


No N/A Yes

Yes

2. o058525860010_01_P001 30.10.2

018 No N/A Yes


Yes


No N/A Yes

Yes

3. o058525860010_01_P001 30.10.2

018 No N/A Yes


Yes


No N/A Yes

Yes





No. Product Name

Processing Date

Area Coverage

No

. o

f C

lass

es

Re

fere

nce

Da

ta

Sa

mp

le S

ize

Sa

mp

ling U

nit

(21

)

Sa

mp

ling

De

sign

(22

)

Sa

mp

le E

xclu

sio

n

Cri

teri

a (2

3)

Th

em

ati

c A

ccu

racy

(2

4)

Acc

ep

tan

ce S

tatu

s

Comment

1.

Urban Green Areas 2018


15 VHR image 2018


STRAT N/A 94,3% Yes




5.2 Error Matrices

Class No

Open Green Areas 2018 Totals

1 2 3 4 5 6 7 8 9 1

0 11

12

13 14 15

Reference Data Other

Square –

Neighborhood

Square – Pock

et

Square – suburban

Other

open spac

e

Market

(Open sky)

Waterfront

Park –

Neighborhood

Park – Pocket

Green – Linear

Green – Trees

Green – Other

Cemetery

Green – Sport & Leisure

Inaccessible potential PS

1 Other 55 0 0 0 0 0 0 0 0 0 0 0 0 0 0 55

2 Square – Neighborhood 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 2

3 Square – Pocket 0 0 2 0 0 0 0 0 0 0 0 1 0 0 0 4

4 Square – suburban 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 2

5 Other open space 1 0 0 0 78 0 0 0 0 0 0 3 0 2 0 84

6 Market (Open sky) 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 3

7 Waterfront 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 2

8 Park – Neighborhood 0 0 0 0 0 0 0 14 0 0 0 0 0 0 0 14

9 Park – Pocket 0 0 0 0 0 0 0 0 11 0 0 0 0 0 0 11




10

Green – Linear 0 0 0 0 0 0 0 0 0 9 0 0 0 0 0 9

11

Green – Trees, Forest, Woodland 1 0 0 0 0 0 0 0 0 0 4 0 0 0 0 5

12

Green – Other 0 0 0 0 0 0 0 0 0 0 0 20 0 1 0 21

13

Cemetery 0 0 0 0 0 0 0 0 0 0 0 0 11 0 0 11

14

Green – Sport & Leisure 0 0 0 0 1 0 1 1 0 1 0 0 0 53 0 57

15

Inaccessible potential PS 2 0 0 0 0 0 0 0 0 0 0 0 0 0 1 3

Totals 60 2 3 1 79 3 3 15 11 10 4 24 11 56 1 283

Accuracy Statistics z= 1,96

Overall Accuracy: 94,3%

95% Confiden

ce Interval:

91,65%

97,04%

Kappa (Cohen): 0,85

Class Name Producer Accuracy

95% Confidence Interval

Producer Accuracy

User Accuracy

95% Confidence Interval

Producer Accuracy

Other 100,0% 100,0% 100,0% 91,7% 88,4%

94,9

%

Square – Neighborhood 100,0% 100,0% 100,0% 100,0% 100,0% 100,0%

Square – Pocket 75,0% 70,0% 80,0% 100,0% 100,0% 100,0%

Square – suburban 50,0% 44,2% 55,8% 100,0% 100,0% 100,0%




Other open space 92,9% 89,9% 95,9% 98,7% 97,4% 100,0%

Market (Open sky) 100,0% 100,0% 100,0% 100,0% 100,0% 100,0%

Waterfront 100,0% 100,0% 100,0% 66,7% 61,2% 72,2

%

Park – Neighborhood 100,0% 100,0% 100,0% 93,3% 90,4% 96,2

%

Park – Pocket 100,0% 100,0% 100,0% 100,0% 100,0% 100,0%

Green – Linear 100,0% 100,0% 100,0% 90,0% 86,5% 93,5

%

Green – Trees, Forest, Woodland 80,0% 75,3% 84,7% 100,0% 100,0% 100,0%

Green – Other 95,2% 92,8% 97,7% 83,3% 79,0% 87,7

%

Cemetery 100,0% 100,0% 100,0% 100,0% 100,0% 100,0%

Green – Sport & Leisure 93,0% 90,0% 96,0% 94,6% 92,0% 97,3

%

Inaccessible potential PS 33,3% 27,8% 38,8% 100,0% 100,0% 100,0%




6.1 Completeness

INPUT DATA

No.






Metadata Comments


250km² 100% 3 3 100% Yes none


PRODUCTS

No.







Metadata Comments

1. Open Green Areas 2018

Bamako (250 km2)

250 km² 100% 100% N/A 100% Yes none




6.2 Compliancy

Product 1 (28) Urban Green Areas 2018

Abstract

Urban green areas refer to land within and on the edges of a city that is partly or completely covered with grass, trees, shrubs, or other vegetation. This includes public parks, private gardens, cemeteries, forested areas as well as trees, river alignments, hedges, etc. The product delivered within EO4SD-Urban project thus provides accurate information (1m resolution) on the spatial location and extent of the green areas located within the Urban Extent derived from baseline LULC information product.


Area Coverage








N/A N/A




Most actual 2018


N/A N/A






N/A All classes were mapped according to their definition.

yes




Spatial Resolution









Data Type



Bit Depth



Data Format






Na Na ---

Metadata



product Yes ---


Thematic Accuracy



94,3% Yes

Positional Accuracy



Delivery Procedure

Service Provision



Delivery Date





March 2019




Glossary














Purpose (9)















State whether (and which) radiometric processing was applied (e.g., contrast enhancement, histogram matching, bundle block adjustment, radiometric normalisation, filtering, etc.).

























Lineage (29)













Product: LULC


Sheet filled in

Requirements






Data Quality Checks








Thematic Processing



Accuracy Assessment














0.1 Requirements

Product 1 (28) Transportation network

Abstract

Transportation network in Core City of Bamako, Mali. Extracted from OpenStreetMap layers improved by visual re-interpretation. Harmonization of road classes and update of missing ones.


Area Coverage




Peri-Urban: N/A




Comments:


UTM 29N


Arterial Motorway, express roads as well as some urban ring roads

Collector Roads with an area distributor function allow entering and leaving residential areas.

Local Roads with an access function allow actual access to properties alongside a road or street.

Railway Railways and associated areas.



Spatial Resolution




Data Type & Format





Bit Depth

N/A

Class Coding


N/A N/A N/A

Metadata



Thematic Accuracy


Positional Accuracy

RMSE < 30 m.

Delivery Procedure

Service Provision

Online via FTP

Delivery Date

March 2019.




1.1 List of EO Data


Incoming Date

Acquisition Date

Proc. Level

Path / Row


Spatial Res.

No. of Bands



Data Format (3)

Bit Depth (5)



WorldView-3

1. 058525860010_01_P001 15.10.2018

4.1.2018


4 0% UTM29 / WGS84

*.tif 16 Bit u

*.imd / *.xml

2. 058525860010_01_P002 15.10.2018

14.1.2018


4 0% UTM29 / WGS84

*.tif 16 Bit u

*.imd / *.xml

3. 058525860010_01_P003 15.10.2018

14.1.2018


4 0% UTM29 / WGS84

*.tif 16 Bit u

*.imd / *.xml





N/A


Dataset 1 (14)


Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)



No. of Classes

Thematic Accuracy


SRTM1arcsec 9.10.2018 *.mta 11.-22.02. 2000



N/A unknown N/A

Lineage(29):



Dataset 2 (14)


Metadata (2)



Area Coverage

Data Type (4)

Data Format (3)


(16)


No. of Classes

Thematic Accuracy

Class Definitions



N/A N/A unknown Yes







2.1 EO Data Quality


Ba

cku

p

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/

Me

tad

ata

(2)

Band Specifications

Pro

ject

ion

/ S

ph

ero

id (1

6)

Sce

ne

Lo

cati

on

Co

mp

lete

ne

ss o

f A

dd

itio

na

l D

ata

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Clo

ud

Co

ver

(in

tern

al

che

ck)

Ra

dio

me

try

To

po

gra

ph

y

Dro

pp

ed

Lin

es

/ A

rte

fact

s (7

)

Acc

ep

tan

ce S

tatu

s

Comment


No

. o

f B

an

ds

Ba

nd

Re

gis

tra

tio

n

Sp

ect

ral/

Sp

ati

al

Re

solu

tio

n

WorldView-3

1. 058525860010_01_P001 Y - Q

Y - V Q

Y - V Q

Y - V Q

Y - V Q

Y - Q

Y - V Y - V Y - V

Q Y - Q

Y - Q

Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

2. 058525860010_01_P002 Y - Q

Y - V Q

Y - V Q

Y - V Q

Y - V Q

Y - Q

Y - V Y - V Y - V

Q Y - Q

Y - Q

Y - V Q

Y - V Q

Y - V

N - V Q

Yes none

3. 058525860010_01_P003 Y - Q

Y - V Q

Y - V Q

Y - V Q

Y - V Q

Y - Q

Y - V Y - V Y - V

Q Y - Q

Y - Q

Y - V Q

Y - V Q

Y - V

N - V Q

Yes none





N/A





Dataset 1 (14) DEM

Re

ad

ab

ility

(1

)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

M

atc

he

s S

erv

ice

Are

a (

%)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

a

nd

un

it (

e.g

. m

, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/ A

rte

fact

s (

EO

da

ta

on

ly)(

7)

Uti

lity

for

Cu

rre

nt

Pro

ject


SRTM1arcsec ☒ Yes

☐ No


☐ Incomplete

☒ not available

☒ Yes

☐ No


☒ Correct

☐ Incorrect



☒ Complete

☐Incomplete

☒ No

☐ Yes


☒ No

☐ Yes


☐ None

☐ Partial

☒ Full


Dataset 2 (14) OSM

Re

ad

ab

ility

(1)

Ch

eck

He

ad

er

/Me

tad

ata

(2)

Da

ta C

ove

rage

Ma

tch

es

Se

rvic

e

Are

a (

%)

Pro

ject

ion

/ S

ph

ero

id

, E

PS

G(1

6)

Sp

ati

al R

eso

luti

on

an

d u

nit

(e

.g.

m, k

m)

Da

ta F

orm

at

(3)

Bit

De

pth

(5)

Co

mp

lete

ne

ss

(Ve

cto

r: A

ttri

bu

te

Ta

ble

; R

ast

er:

Th

em

ati

c V

alu

es)

Ge

om

. M

isa

lign

me

nt

Dro

pp

ed

Lin

es

/ A

rte

fact

s (

EO

da

ta

on

ly)(

7)

Uti

lity

for

Cu

rre

nt

Pro

ject



☐ No


☐ Incomplete

☒ not available

☒ Yes

☐ No


☒ Correct

☐ Incorrect


n.a. *.shp n.a.

☒ Complete

☐Incomplete

☒ No

☐ Yes


☒ No

☐ Yes


☐ None

☐ Partial

☒ Full

Comments:






Pro

cess

ing D

ate

AO

I

Cit

y /

Re

gio

n /

Co

un

try

Pro

ject

ion

/ S

ph

ero

id (1

6)

No

. &

RM

S (

m)

of

GC

Ps

(17

)

No

. &

RM

S (

m)

of

TP

s (1

8)

No

. &

RM

S (

m)

of

CP

s (1

9)

Dig

ita

l E

leva

tio

n M

od

el (D

EM

)

Mo

de

l /

Alg

ori

thm

(13

)

Re

sam

plin

g

Me

tho

d (2

0)

Va

lida

tio

n R

ep

ort

s

Acc

ep

tan

ce S

tatu

s



WorldView-3

1. 058525860010_01_P001 17.10.2018

Bamako

UTM29 / WGS84

9; X RMSE 0.66, Y RMSE 0.62

N/A N/A SRTM

30


CUB

Yes Yes o058525860010_01_

P001

2. 058525860010_01_P002 17.10.2018

Bamako

UTM29 / WGS84

9; X RMSE 0.78, Y RMSE 0.37

N/A N/A SRTM

30


CUB

Yes Yes o058525860010_01_

P001

3. 058525860010_01_P003 17.10.2018

Bamako

UTM29 / WGS84

10; X RMSE 0.56, Y RMSE 0.46

N/A N/A SRTM

30


CUB

Yes Yes o058525860010_01_

P001




3.1.1 Data Fusion


Acc

ep

tan

ce S

tatu

s

Comments

Filename 058525860010_01_P001_M

UL 058525860010_01_P001_PA

N 058525860010_01_P001_PS






Data Format (3) &


16 Bit u

Yes


Acc

ep

tan

ce S

tatu

s

Comments

Filename 058525860010_01_P002_M

UL 058525860010_01_P002_PA

N 058525860010_01_P002_PS






Data Format (3) &


16 Bit u

Yes


Acc

ep

tan

ce S

tatu

s

Comments

Filename 058525860010_01_P003_M

UL 058525860010_01_P003_PA

N 058525860010_01_P003_PS






Data Format (3) &


16 Bit u

Yes




3.2 Data Processing




OTHER Calculation:

Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment

No. Processing

Date

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software / Method

WorldView-3







4.1 Classification


Processing

Date

Cloud Masking


Manual Enhancement

Border Matching

Acc

ep

tan

ce S

tatu

s

Ba

cku

p

Comment N


pro

cess

ed

Software /

Method p

roce

sse

d

Software / Method

pro

cess

ed

Software / Method

pro

cess

ed

Software /

Method

WorldView-3

1. o058525860010_01_P001 30.10.2

018 No N/A Yes


Yes


No N/A Yes

Yes

2. o058525860010_01_P001 30.10.2

018 No N/A Yes


Yes


No N/A Yes

Yes

3. o058525860010_01_P001 30.10.2

018 No N/A Yes


Yes


No N/A Yes

Yes





No. Product Name Processing Date

Area Coverage

No

. o

f C

lass

es

Re

fere

nce

Da

ta

Sa

mp

le S

ize

Sa

mp

ling U

nit

(21

)

Sa

mp

ling

De

sign

(22

)

Sa

mp

le E

xclu

sio

n

Cri

teri

a (2

3)

Th

em

ati

c A

ccu

racy

(2

4)

Acc

ep

tan

ce S

tatu

s

Comment

1. Transportation network 2018


N/A VHR image 2018


STRAT N/A 94,6% Yes




6.1 Completeness

INPUT DATA

No.






Metadata Comments


250km² 100% 3 3 100% Yes none


PRODUCTS

No.







Metadata Comments

1. Transportation network 2018

Bamako (250 km2)

250 km² 100% 100% N/A 100% Yes none




6.2 Compliancy

Product 1 (28) Transportation network

Abstract

Transportation network in Core City of Ramadi, Iraq. Extracted from OpenStreetMap layers improved by visual re-interpretation. Harmonization of road classes and update of missing ones.


Area Coverage








N/A N/A




Most actual 2018


N/A N/A






N/A All classes were mapped according to their definition.

yes




Spatial Resolution









Data Type



Bit Depth



Data Format






Na Na ---

Metadata



product Yes ---


Thematic Accuracy



94,6% Yes

Positional Accuracy



Delivery Procedure

Service Provision



Delivery Date


March 2019

Glossary

















Purpose (9)











Radiometric Processing (15) State whether (and which) radiometric processing was applied (e.g., contrast enhancement, histogram matching, bundle block adjustment, radiometric normalisation, filtering, etc.).




























Lineage (29)









Page 14


Date post:	22-Aug-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times