Manual de ERDAS

Copyright 2010 ERDAS, Inc. All Rights Reserved.Printed in the United States of America.

cc02/06 Part No. 1873

The information contained in this document is the exclusive property of ERDAS, Inc. This work is protected under United States copyright law and other international copyright treaties and conventions, to include the Berne and Geneva Phonograms Conventions, the WIPO Copyright Treaty, and the World Trade Organization.

No part of this work may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, or by any information storage or retrieval system, except as expressly permitted in writing by ERDAS, Inc. All requests should be sent to the attention of Manager of Customer Education at the following address:

ERDAS, Inc.5051 Peachtree Corners CircleNorcross, Georgia 30092-2500 USA

Phone: 770 / 776-3400Fax: 770 / 776-3694

Support Services: 800 / 953-6643Customer Education: 800 / 373-2713Web: www.erdas.com

The information contained in this document is subject to change without notice.

Warning

All information in this document, as well as the software to which it pertains, is proprietary material of ERDAS, Inc., and is subject to a ERDAS license and non-disclosure agreement. Neither the software nor the documentation may be reproduced in any manner, without the prior written permission of ERDAS. Specifications are subject to change without notice.

About This Manual

This Introduction to LPS exercise manual is an instructional document designed to be a part of the Intermediate level courses offered by ERDAS Educa-tion Services. At the Intermediate level, the courses offered consist of the IMAGINE VirtualGIS, Using Imagery to Update Your GIS, Cartography with Map Composer, and the Introduction to the LPS course. All ERDAS education courses provide corresponding instructional documentation.

Trademarks

ERDAS IMAGINE, IMAGINE Essentials, IMAGINE Advantage, IMAGINE Professional, IMAGINE GLT, TopoMouse, and Stereo Analyst are registered trademarks of ERDAS. LPS, IMAGINE VirtualGIS, and CellArray are trademarks of ERDAS. Geospatial Imaging Chain is a service mark of ERDAS, Inc. Other brands and product names are trademarks of their respective owners.

Acknowledgments

The QuickBird imagery and associated data provided in this Photogrammetry Course and Exercise Manual are courtesy of DigitalGlobe, Longmont, CO. Any use or distribution of this data by others, or outside the realm of ERDAS Education Services, is strictly prohibited.

iv Introduction to LPS

ERDAS Customer Education Introduction to LPS 1

Table of ContentsTable of Contents .........................................................................1

Using This Manual .........................................................................3

Section 1: Creating a BlockExercise 1: Raw Scanned Imagery .................................................... 11

Challenge 1: Scale Variation .......................................................... 19

Exercise 2: LPS Setup Wizard ......................................................... 25

Exercise 3: Imagery Requirements ................................................... 35

Exercise 4: Inside the Sensor.......................................................... 41

Exercise 5: The Sensor in Space ...................................................... 51

Exercise 6: Take Control ............................................................... 55

Exercise 7: Building the Relationship ................................................ 83

Exercise 8: The Triangulation Results ............................................... 91

Section 2: DTM ExtractionExercise 1: In Search of Z.............................................................. 97

Exercise 2: Visualizing Your DEM ....................................................109

Exercise 3: Statistical DTM Evaluation..............................................113

Exercise 4: Verifying Results .........................................................117

Section 3: Creating and Working with Ortho ImagesExercise 1: Orthorectification .......................................................125

Exercise 2: Working with DigitalGlobe Data .......................................129

Exercise 3: Mosaic Images ............................................................147

Exercise 4: Visualizing Your Images .................................................153

2 Introduction to LPS Table of Contents

Section 4: LPS Terrain EditorExercise 1: Introduction...............................................................161

Exercise 2: DTM Editing Tools ........................................................177

Exercise 3: Setting a Constant Elevation ...........................................189

AppendixAppendix A - Geometric Models in LPS .............................................197

Appendix B - Terrain Editor Operators .............................................201


Using This Manual

Introduction

This manual contains step-by-step instructions on how to perform certain processes. You should be aware that each exercise provides a single path through ERDAS IMAGINE and LPS tools. In most cases, there are various ways to maximize tool usage, depending on the individual project.

This exercise manual is provided to the student, along with all images and copies of the presentation slides used by the instructor. This provides the capability for recreating the processes performed in class at a later date, as well as the key points on any theory involved.

Exercise Conventions

Section Title Page

States the objective of the exercises and lists the application tools to be utilized within the various tasks.

Exercise Tasks Each exercise is split into a set of tasks. After the course, these tasks will help you locate within the manual where you performed a certain set of steps.

Questions These appear in a larger font with both bold and italic attributes. The instructor may quiz and/or review with you, following each exercise.

Class Notes These pages follow the end of each exercise on which notes from your studies can be made.

4 Introduction to LPS Using This Manual

Notational Conventions

Bold Italicized Text

Any text, which is bold, indicates a filename, or parameter to be changed or selected.

Graphics To help you locate icons and objects used in the exercises, the icons will be next to the icon name in the text.

Insets These italicized captions will appear in the outside margin of the page. They define terms or explain theory behind the steps you are being asked to perform.

Diagrams These are an optional means to direct you in the usage of some of the applica-tions tools.

The following graphics are also used for particular purposes:

This is a note or a quick tip. It gives additional relevant infor-mation, or describes other ways of using the software.

This is information relating to the application of the current tools.

This is a reference. It provides additional theory or science that will help in using the tools.

This is a warning. It cautions you regarding potential pitfalls and how to avoid producing errors.


Mouse and Keyboard Conventions

LMB Left mouse button

MMB Middle mouse button

RMB Right mouse button

LMB + Shift Hold down the Shift key plus left mouse button


Windows Terminology

Title bar

Open file button

Radio button (enabled)

Radio button (disabled)

Text box with nudgers

Popup list

Checkbox (enabled)

Button (enabled)

Button (disabled)


The eWorkspace

ERDAS Application Menu button

QuickAccessToolbar Title bar

Active Layer type

Hiddenpanel

Dockedpanel

Ribbon

2D View Status BarMap View


The Ribbon Explained

Group

CollapsedGroup

ExpandedGroup

Unselected tab

Properties button

Selected tab


Section 1: Creating a Block

Section Objective

Students will use the ERDAS IMAGINE Viewer to study scanned images as well as examine the inherent errors in raw imagery. This will help explain the need for orthorectification.

Subsequent exercises will guide you through the process of setting up LPS, entering information about the sensor used, and adding control points.

Finally, the section will explain the triangulation process and how to evaluate the results of the process.

Tools Used

Viewer Used to visualize imagery, zooming, inquiring and measuring features.

LPS Set Up Wizard

Set up the basics parameters associated with a block.

Point Measurement Tool

Used to digitize and measure control and tie points on your imagery.

Triangulation Report

View the statistics and residuals of a triangulation to evaluate accuracy and help locate errors.

10 Introduction to LPS Section 1: Creating a Block


Exercise 1: Raw Scanned Imagery

Objective: To become familiar with the dataset and review the eWorkspace.

Task 1: Setting Session Prefer-ences

1. Start ERDAS IMAGINE.

The ERDAS IMAGINE eWorkspace opens.

1

3

5

6

7

8

4

2


2. From the eWorkspace, click the ERDAS Application button . From the menu, select Open > Raster Layer. The Select Layer to Add dialog displays. This dialog is used throughout ERDAS IMAGINE.

3. From the ERDAS IMAGINE Main Menu select Session | Prefer-ences. The Preference Editor displays.

1) ERDAS Application Menu button

Access to New, Open, Save, View, Print, Session, Batch, Configuration, Preferences and Help.

2) Quick Access Menu One-click access to commonly used functions. Customizable.

3) Ribbon Access to IMAGINE functions, collected in tabs and groups

4) Title bar Window Title, Window (Viewer) display icons: Minimize, Maximize, Close

5) 2D View Main Viewing space. Can add multiple views, Map View, 3D View.

6) Contents Pane Display and arrange all layers each View.

7) Shoebox Organizes easily-accessed shortcuts to your data.

8) Status Bar Cursor identification and image coordinates


4. From the Category list, ensure that User Interface & Session is selected. This displays the default preferences that are associated with the data location and data properties.

5. As the Default Data Directory, type the path to where the course data is located. (Your instructor will provide this information.)

6. As the Default Output Directory, type the path to where all outputs generated in this course will be saved. (Your instructor will provide this information.)

7. In the Category list, select Viewer, and the preferences for the Viewer display.

8. Scroll down (vertical scroll bar on the right side) to the Clear Display, Fit to Frame, and Background Transparent options.

By default, the Clear Display option is enabled and the Fit to Frame and Back-ground Transparent options are disabled.

9. Enable the Fit to Frame and Background Transparent checkboxes and disable the Clear Display checkbox to set these as new defaults for the Raster Options dialog.

10. Click User Save and Close.


Task 2: Display a Raw Scanned Aerial Photo

In this exercise we will be using the eWorkspace to view an image.

1. From the eWorkspace, click the ERDAS Application button . From the menu, select Open > Raster Layer. The Select Layer to Add dialog displays.

2. In the Select Layer to Add, navigate to the LPS directory, and ensure the Files of type is set to IMAGINE Image (*.img), then select the file: 11_189.img

3. Before clicking Ok, click the Raster Options tab, ensure that the Fit to Frame checkbox is enabled, and for the Layers to Colors, set Red to 1, Green to 2 and Blue to 3, then click OK.

4. In the Attention dialog, disable the Always Ask checkbox and click Yes.

Task 3: Locating and Identifying Fea-tures

In the next steps, you will identify a location on the image using the Inquire Cursor.

1. Right Mouse Button (RMB) click on the image and from the Quick View menu, select Inquire Cursor. This will start the Inquire Cursor and display the Inquire Cursor dialog.

2. Move the Inquire Cursor onto the image.

We are calculating Pyramid Layers to facilitate quick view-ing within IMAGINE. LPS pyramid layers are more complex than IMAGINE pyramid layers, so we will ahve to recalcu-late them again later.

You may also select the Inquire Cursor by clicking on its

icon from the Viewer toolbar.


3. In the Inquire Cursor dialog, click the pull-down arrow for Coordi-nate Type and select File.

What are the units of measurement the Coordinate Type-File is using?

Where on the image is the origin of the file coordinate type?

4. Input the following X and Y coordinates, then press Enter:

While viewing a feature, there are a number of tools that let you zoom in and out. The tools used in this exercise are the Interactive Zoom In and the Interactive Zoom Out.

5. Click the Interactive Zoom In icon .

6. Hold down the Left Mouse Button (LMB) and drag the mouse to place a box around the feature identified above. When you release the mouse button, the view will zoom in.

7. To move around the image, click the Roam Image icon , then LMB click and hold to drag the image.

X = 3579.80 Y= 2634.30

The pointer displays as a magnifying glass as you posi-tion it on the image.

The Interactive Zoom In icon can be used to zoom in 2 times by left clicking on the image.

The Interactive Zoom Out icon can be used to zoom out 2 times by clicking on the image.


8. Take a minute to practice roaming around the image, looking at differ-ent features found in Denver.

9. Click Close in the Inquire Cursor dialog.

10. Close the Viewer without saving changes.


Class Notes


Challenge 1: Scale Variation

Objective: To identify inconsistencies with are inherent in raw scanned imagery

Task 1: Scale Calculation Using Raw Imag-ery

On these raw images, it is difficult to see any scale variation. To help us check for scale variations throughout the raw image, we will compare measurements at two different photo locations.

Ground distances collected from other sources will be used to help us determine photo scales.

Scale Variation: occurs in all photogra-phy, due to objects being closer to or fur-ther away from the camera. The scale is not constant across a photo.

The first distance you will measure is the length of the field at the coordinates listed previously.

1. In the eWorkspace, click the Open Layer icon on the Quick Access toolbar.

2. In the Select Layer to Add, navigate to the LPS directory, and ensure the Files of type is set to IMAGINE Image (*.img), then select the file: 11_189.img. The image may take a minute to load.

3. Use the Zoom and Roam tools so that you can see the length of the field in your Viewer.

4. From the Home tab, Information group, click the Measure icon .

The Measurement Tool displays.


5. Leave the Measurement Units set to the default (Other), which rep-resents pixels.

6. Use the Zoom tools to zoom into the bottom left end of the field.


7. From the Measurement Tool dialog, click the Measure Lengths and Angles icon .

8. Place the cursor over this point and click to take the beginning mea-surement. You should now see a line extending from this point to your cursor.

9. Roam to the other end of the field by moving your cursor outside the Viewer. Continue roaming until you reach the identified location in the images, then place the cursor over that point and double-click to take your end measurement.

10. Make a note of the distance displayed: ______________ pixels

11. Use the following equation to calculate the distance in centimeters:

__________ pixels / 400 = __________ centimeters

12. Then note your results:

Football field length is ______________ centimeters

To scroll through the image horizontally, use Shift + Scroll-wheel.

To scroll vertically through the image, use Ctrl + Scroll-wheel.

To change pixel measurements into centimeters, we must know the scanning resolution. These images were scanned at 25 microns, meaning that each pixel is 25 microns across. 10,000 microns equals 1 centimeter. Therefore, there are 400 pixels in a centimeter.


13. Click Zoom to Data Extent .

14. Open the Inquire Cursor and navigate to following coordinates:

X= 5493 Y= - 6943

15. Zoom in on the pitchers mound on which the Inquire Cursor is placed.

16. From the Measurement Tool dialog, click the Measure Lengths and Angles icon .

The distance you will be measuring is defined by a line in the image following. The line extends from the baseball pitchers mound to the football 50 yard line.


17. Place the cursor over the pitchers mound and click to take the begin-ning measurement.

18. RMB click in the Viewer. This temporarily turns off the measure-ment that you are taking.

19. In the Inquire Cursor dialog, input the following coordinates:

X= 3576 Y= - 2677

20. Move the cursor back into the Viewer and the measurement will be on again. Double-click on the 50 yard line of the football field.

What is the length you measured? ______________ centimeters

From these raw photo distances, the photo scale can be determined. Listed below are the ground distances for the above features.

What are some sources where we can get ground distances?

21. Calculate the photo scale using the following equation:

Photo scale = (ground distance * 100) / photo distance

22. Record the scale for each of the measured distances:

What causes the scale variations?

If there were no changes, what would this mean in reference to the orientation of the airplane and the ground?

23. Close all open Viewers and tools without saving changes.

Football Field 91.44 meters

Pitcher's mound to 50 yard line 2608.10 meters

Football Field 1: __________

Pitcher's mound to 50 yard line 1: __________


Exercise 2: LPS Setup Wizard

Objective: To open a new LPS project and input the setup parameters required by the Block Properties Setup tool.

You will begin your work with a Setup Wizard that guides you through the process of defining the project. The linear workflow helps to quickly select your camera or sensor model, the projection, spheroid and datum, and the units.

Task 1: Defin-ing a Project

1. From the Toolbox tab, click the LPS icon . This will launch LPS without a project loaded.

2. From the LPS menu, select File | New. The Create New Block File dialog displays.

3. Navigate to the Outputs directory and type denver_frame.blk as the File name. Click OK.

Alternatively, you can click the Create new block file icon .


Task 2: Choosing a Camera Model and Map Projection

1. From the Model Setup dialog, select Frame Camera, then click OK. The Block Property Setup box displays. From this, a reference sys-tem must be set. In a production environment it may take time to research the reference system.

This dialog is used to select the model for use with your block file images. The model defines the geometric proper-ties associated with the sensor or camera.

All Geometric Models are listed and described in Appendix A.


2. Click the Set button. The Projection Chooser dialog displays.

A few parameters must be set in order to define the Map Projection.

3. Display denver-5m-dem.img in a Viewer.

4. To obtain the projection, select the Home tab and then click the Layer

Info button . The ImageInfo dialog displays.

5. Click the Projection tab.

The Projection Chooser handles the selection of LPS's standard projections, or creates a custom projection.

During this exercise a custom projection will be created, and saved as a standard projection.


6. Fill in the following information:

7. In the ImageInfo dialog, select File | Close, then close the Viewer.

8. In the Projection Chooser dialog, click the Custom tab, and in the appropriate fields, input the parameters collected above, then click Save.

Is the data North or South of the equator?

9. For the new Projection Name, type Denver and for the In Category, select UTM WGS 84 North, then click OK.

What are some circumstances, which require you to create more than one block for use in the same project?

Projection _______________

Spheroid ________________

Datum ___________________

Zone Number _____________

North or South ____________

It is important to enter the units correctly. Once the block has been created there is no going back to change the units. A new block would have to be created and valuable time may be lost.

The information does not have to be typed into the boxes. You may choose the appropriate information from the pull-down menus.

By saving Denver to the Standard projection tab, the new projection can be used at a later date.


10. Click the Standard tab, and in the UTM WGS 84 North category, select Denver as your projection, then click OK. The projection infor-mation displays in the Block Property Setup.

11. Click the Next button.

12. For this exercise, set the Angle Units to Degrees.

Task 3: Setting Frame-Specific Information

The Rotation System and the Photo Direction must be set. Each image or photo is defined in its relation to the ground, by 3 rotation angles (omega, phi, kappa).

In the most common rotation system, Omega is the primary axis. The primary axis is the axis about which the first rotation occurs. Positive rotations occur clockwise around the axes, as you look in a positive direction along the axis (e.g., looking from 0 to X).

In the Denver data, around which axes are the following rotations? (Use On-Line Help as a reference)

These photos of Denver use the Omega, Phi, Kappa Rota-tion System (Position and orientation of the sensor as it existed when the image or photo was captured).

Omega = ___________ axis

Phi = ___________ axis

Kappa = ___________ axis


Which parts of the world use the following Rotation Systems?

13. We will leave the Photo Direction as the default, Z-axis for normal images.

For what type of photography would you set the Photo Direction to the Y-axis?

14. Using the following equation, calculate the flying height, then input the result into the appropriate box in the Block Properties Setup.

Phi(+), Omega, Kappa __________________

Phi(-), Omega, Kappa ___________________

The Photo Direction must also be set. The Photo Direction equals the Optical Axis of the Camera. The Optical Axis is a straight line extending from the perspective center, through the center of the lens to the ground. The perspective center is the point in space where the image was exposed.

The Import Exterior Orientation Parameters button allows for the import of existing exterior orientation data, that are provided by photogrammetric systems, such as analytical stereo plotters, digital photogrammetric workstations, or post-processed airborne GPS data recorded at the time of image capture.


Flying Height (fh) = Focal Length (f) x Scale (h) / 1000

15. Click OK to close the Block Property Setup dialog, which automati-cally opens the LPS dialog.

16. From the LPS menu, select File | Save. It is good practice to save your block before continuing.

What are some of the parameters set during the Block setup?

17. After the Block has been saved, select File | Close.

The block you have just created is based on a Frame Camera Geometric Model. The block can be reopened at a later time.

Task 4: SPOT Push-broom Geometric Model

For this next task, you will be creating another block but this time, it will be based on a SPOT Pushbroom model.

1. Click the Create new block file icon .

2. From the dialog that displays, click the Goto button, and navigate to the Outputs directory.

3. For the block File name, type denver_spot, then press Enter.

Polynomial-based Pushbroom: A sen-sor type which scans an area along parallel lines, perpendicular to the direction of move-ment.

4. Click OK and the Model Setup dialog displays. From the dialog, select Polynomial-based Pushbroom from the Geometric Model Category.

5. As the Geometric Model, select SPOT Pushbroom, then click OK.

6. Click Set. The Projection Chooser dialog displays.

The photography for Denver was taken with a 153.475 mm focal length camera and the photo scale in approximately 1:21,000.

Make sure all your parameters have been entered correctly before you click OK. Once acknowledged, the block is defined and only the flying height can later be altered (Edit | Block Properties | Model-Specific). If changes thereafter, are necessary, a new block will have to be defined.


7. From the Standard tab, select the Denver projection you created previously, then click OK.

8. Once the projection reference has been set, ensure that Meters has been selected as the Horizontal Units.

What is the major difference you have noted between setting up a Frame Camera and a Spot Geometric Model?

9. In the Block Properties Setup, click OK. The block is created and the project is now listed in the LPS Project Manager.

10. To save the block, select File | Save, then from the LPS Block Tool, select File | Close.


Class Notes


Exercise 3: Imagery Requirements

Objective: Review the imagery formats, which can be imported into LPS, and generate pyra-mid layers.

Our block has been setup and defined, the main LPS dialog displays, and the task now, is to bring in the imagery. Before processing, we must understand the imag-ery requirements and limitations.

Task 1: Import Files into the Block

The LPS Project Manager window is displayed but does not have a project loaded.


1. Click the Open block file icon . The Block File Name dialog dis-plays

2. Click the Recent button and select denver_frame.blk, and then click OK.

3. In the Block File Name dialog, click OK.

4. From the LPS icon panel, click the Add Frame to the List icon . The Image File Name dialog displays.

Dynamic Loadable Library (DLL): Allows non-native LPS files to be read directly by the application.

5. Use the Files of Type pull-down menu to look at the file types avail-able for loading into LPS.


The list displays file formats, which use a DLL when being loaded for viewing by LPS.

6. From the Image File Name dialog, select 11_188.img, 11_189.img, and 11_190.img. Click OK.

7. Update the Description column by giving 11_188.img the name denver1. Give 11_189.img the name denver2 and 11_190.img the name denver3.

8. Take a look at the column headings and the contents of the columns.

When inputting images, you have the option of inputting one image at a time, or a number of images.

This is accomplished by selecting one image at a time or by highlighting the first image, holding the SHIFT key and clicking on the last image to add, then clicking OK to load the data.

If you are unable to see the full Description, place the mouse cursor to the right of the word Description. The cur-sor changes to a dual-ended arrow.

Pyramid Layers

Interior Orientation

Exterior Orientation

Digital Terrain Model

Orthorectified


What do the 'X's in the active column mean, and what processes are associated with this column? (Hint: Use On-Line Help.)

In what situation would the Online column be colored red?

Task 2: Creating Pyramid Layers

Pyramid layers are copies of the original layer(s) that have been successively reduced by a power of 2, and then resampled. The number of pyramid layers cre-ated will depend on the size of the image. A large image will produce more pyra-mid layers.

Pyramid layers are added as additional layers in the .img file. However, these lay-ers cannot be accessed for display.

1. Click on any red-colored cell in the Pyr. column. The Compute Pyra-mid Layers dialog displays, from which you have the following three options:

One Image Selected

All Selected Images

All Images Without Pyramids

2. Maintain the default of All Images Without Pyramids, then click OK. The cells in the pyramid layers should turn green.

The green color patches in the matrix cells indicate that the feature has been created, or that the process has been run and completed. Red indicates the step still needs to be completed.

This visual guide can help you determine which steps have been performed for each of the images in your Block File.

This will take some time depending on the number of images in the list.


What are the differences between ERDAS IMAGINEs pyramid layers and LPS's pyramid layers?

3. In the LPS dialog, select File | Save to save your image list.

We have imported our imagery into an LPS block and created pyramid layers. We must now give LPS the sensor information, and setup the camera, as well as the image parameters.


Exercise 4: Inside the Sensor

Objective: Perform Interior Orientation by entering camera parameters and measuring the Fiducial Marks.

Task 1: Defin-ing the Sensor

At this point, the pyramid layers have been created for the images to be used; the green color patch in the Pyr. column is a visual indication of this.

1. From the LPS menu, select Edit | Frame Editor, or click the Show

and Edit Frame Properties icon .

The Frame Editor dialog displays with the information for the currently selected image (i.e., the image with the caret symbol next to it).

2. To ensure the image you selected is the correct one, click on the View Image button. A Viewer displays your image.

3. Once you have confirmed it is the correct image, close the open Viewer.

The Next and Previous buttons can be used to scroll through all of your images and change parameters, without having to return to the main Block Tool dialog.


ISCS: Many aspects of aerial cameras are accurately measured in a laboratory. These measurements should be provided to you as a report.

4. Click the New Camera button and the Camera Information dialog dis-plays. In this case, the Camera Name refers to the aerial camera type.

5. In the Camera Information dialog, click the Load button, locate and select (highlight) the denver_frame.cam file, then click OK.

6. Review the information from the camera file, such as:

Photogrammetry is essentially the process of establishing a relationship between the camera/sensor, the imagery, and the ground.

To build this mathematical relationship, we must be able to define points on the images, points on the ground, and the location of the sensor.

To define the points on the images, we need to setup an Image-Space Coordinate System (ISCS).

To do this we need information about the camera/sensor itself.

Focal Length

Principal Point xo

Principal Point yo


Fiducials Marks: marks within the cam-era body which are transferred to the film during exposure. The locations of these marks are measured in a laboratory and then placed into the calibra-tion report.

7. Fiducial marks can be located and displayed in the imagery. Click the Fiducials tab.

If you needed to change the Number of Fiducials you could make that change under this tab.

Where on the image are the fiducials located?

8. Click the Radial Lens Distortion tab.

Are the Distortion Coefficients editable?

9. Click OK. The camera information for the first image has been entered.

10. Click Next to ensure that denver_frame is being used as the Sensor Name for image 11_189.img, then repeat for the final image.

Task 2: Fiducial Orientation and Measuring Fiducials

LPS assumes the same camera collected all of the images, so you do not need to input this information for each image. However, if the camera did change, we could use the Next or Previous buttons to navigate to the desired image, and modify the camera and fiducial information.

1. In the Frame Editor dialog, click the Interior Orientation tab. The Film X and Y positions, which were entered by the camera file, are displayed.

What values do the Image X and Image Y columns contain and why are they initially empty?

Lenses are never perfect. The Radial Lens Distortion parameter will model any errors found in the lens by the lab-oratory technician. These errors can then be taken into account during processing.

Although not a requirement, we can add the radial lens distortion to help further define the camera. In this exam-ple, we will not be using any Lens Distortion Parameters.

If the Field Angles and Distortion are entered into the CellArray, the Calculate Coeffs button would be clicked. The three Konrad Distortion Coefficients would then be cal-culated. These are used to mathematically model the distor-tion in the lens.


You will select the correct Fiducial Orientation. To determine the correct orienta-tion, typically, you would locate the data strip. The Y-axis should lie parallel to this strip. In the airphotos provided, the data strip has been removed. Your instructor will indicate to you the orientation of the fiducials.

2. For this exercise, select the fourth icon . This icon best portrays our images orientation.

3. Now, use the Next and Previous buttons to set the Fiducial Orienta-tion for the other two images. Then use the Previous button to return to the first image.

During scanning, how important is it to keep the scan orientation constant?

4. Under the View Fiducial Locator heading, click the Open Viewer icon . A Main View opens on top of the Frame Editor dialog.

5. To prevent an unwanted measurement from being taken, click the Select Image Fiducial icon .

Once the camera parameters have been entered, the fiducial marks must be measured. Their measured pixel location (c,r) is compared to their calibrated coordinates (mm). This allows the position of the principal point of each image to be calculated, which then becomes the origin of the image coordinate system.

The pixel coordinates and the calibrated coordinates are used to calculate transformation coefficients. Any pixel coordinate can then be transformed into image coordinates.


A Main View, an Over View, and a Detail View display. You will use these views to measure the fiducials.

6. Take a minute to select each of the Fiducial Orientation buttons and note the location, (i.e. which corner), of the Link Cursor in the Over View. When finished, return to the correct orientation.

Any of the three views can be used for measuring the fidu-cials; however, it is advisable to measure in the Detail View.

Before measuring the fiducials, verify the positioning of the Link Cursor, as different Fiducial Orientation buttons are selected.

The Link Cursor in the Main View identifies the approximate area of the fiducial. This area is magnified in the Detail View.


7. Click and drag in the center of the Link Cursor in the Main View, and move it over the top of the first fiducial mark.

What effect does this change have on the Detail View?

Some fiducial centers may be difficult to see. Adjustments may be made using the brightness and contrast sliding bar.

8. Place the cursor over the Detail View, RMB click, and select Set Res-ampling Method. The default resampling method Nearest Neighbor displays in the Set Resampling Method dialog.

9. Click OK. The image becomes pixelated.

10. RMB click on the Detail View, select Set Resampling Method then use the pull-down menu to choose Bilinear Interpolation.

11. Repeat this process but for the Resampling Method, choose Cubic Convolution.

Which resampling method best displays the fiducial center?

Once you can accurately identify the center of the fiducial, you are ready to take a measurement.

12. Click the Place Image Fiducial icon . This changes your cursor into a crosshair when placed over any one of the Viewers.

13. Take your first measurement by clicking over the fiducial center in the Detail View.

14. Repeat the previous steps to measure the second fiducial mark.

15. When these first two fiducials have been measured, click the Auto Locate button. The Automatic Interior Orientation window displays, which will be used to locate the remaining fiducial marks.

16. Click the Current Frame radio button, and then click Run.

The fiducial point is measured in Image pixel coordinates, and the Frame Editor CellArray is automatically updated. The display automatically updates and moves to the next fiducial since the Set Automatic Move icon is enabled.

If the Set Automatic Center icon is enabled, then the fiducial that you digitize will snap to the fiducial center if it is within two pixels.


17. When finished, click Accept then Close.

What could cause a higher than expected RMS error?

What is a good RMSE?

18. Click in the row for the Point # you wish to view. Click the Select Image icon if you need to reposition the mark.

19. When you are satisfied with the fiducial measurements in the first image, click the Auto Locate button again, then click the Unsolved Frames radio button.

20. Click Run, and when complete, click Accept then click Close, and inspect the fiducial measurements for the last two frames, just as you did the first.

Your Root Mean Square Error (RMSE), in pixels and microns, is displayed above the Solve button on the Interior Orientation tab of the Frame Editor.

RMSE - After each measurement, each fiducial's file coordi-nates (r,c pixels) are transformed into image coordinates (millimeters). These new coordinates are then retrans-formed back into file coordinates (r,c pixels).

The variation (residual) between the initial file coordinates and the retransformed coordinates, are indications of the closeness-of-fit of the two coordinate sets.

Any RMSE values larger than 0.5 pixels, or ones which are half the scanning resolution of the image, infer systematic and/or measurement errors. Check for film deformation, poor scanning quality, mis-measured points, and incorrect calibration.

Based on your RMSE, you may want to re-measure some of the fiducials. Frequently, the first few measurements are not as accurate as the latter measurements.


21. Once you have finished, click OK in the Frame Camera Frame Editor dialog.

You will notice that the Int. column has turned green; indicating that you are fin-ished defining the interior orientation of the camera.

22. From the LPS Project Manager menu, select File | Save, or click the Save Block Information icon .


Class Notes


Exercise 5: The Sensor in Space

Objective: To input the position and orientation of the sensor (Exterior Orientation).

Task 1: Examine Parameters for Exterior Orienta-tion

You should have your LPS Project Manager open, and it should contain three images, each of which now has the Int. column as green.

1. Open the Frame Editor and click the Exterior Information tab. You will see six columns.

The first three columns define the perspective center coordinates, while the next three represent the rotational angles of this perspective center, and hence, the image.


2. Confirm that the Status row is Unknown for all six elements.

3. LPS will compute:

Can we get the Perspective Center coordinates from the images?

Exterior orientation defines the position and orientation associated with an image.

To build the relationship between Image and Object space, the software needs to determine the rotational differences between the two coordinate systems.

Xo, Yo, and Zo define the location of the perspective center.

Phi (), Omega (), and Kappa () define the rotation of this perspective center.

As part of its calculations Leica Photogrammetry Suite can determine these six (6) values for each image.

The Std. (standard deviation) row represents the reliability you place on the Values. The lower the Std. the more reli-able your coordinate information.

Initial: This status indicates the exterior orientation value is an approximation to the actual value, and will be modified during triangulation.

Fixed: The value will not be modified during triangulation.

Unknown: This status assumes you have no information about the camera's exterior orientation parameters.

Omega () - a rotation around the ______ axisPhi () - a rotation around the __________ axisKappa () - a rotation around the _______ axis


4. Click OK in the Frame Editor.

We are now ready to input control points that form the basis of the mathematical calculations used to tie the image to the ground.

5. In the Project Manager, click the Save icon .

Be aware that measurements and parameters input are not saved into the Block by clicking OK.

Once the exterior parameters are input, LPS uses these values to establish the positional and rotational relationship between the image space coordinate system and the ground space coordinate system.


Exercise 6: Take Control

Objective: You will become familiar with some of the tools used to input and collect ground control points (GCPs). After measuring the fiducials of each image in your block, you are ready to measure the position of points on the ground.

Task 1: The Point Measure-ment Tool

1. From the LPS menu, click the Point Measurement icon .

2. Ensure that the Classic Point Measurement Tool radio button is selected.

3. Click OK. The Point Measurement dialog displays, which is divided into sections as labeled on the following page.


In the Right View and the Left View of the Point Measurement dialog, there are options to change or specify the image displayed.

To build the relationship between the ground and images, we need to share some values measured in each space coordinate system (Object & Image).

For the ground or object space, we measure GCPs using a coordinate system (X,Y,Z). These same points are then measured on the images (x,y). These two sets of numbers are then used to start to build this relationship.

Point Measurement Group

Left View Group Right View Group Tools Group

Over View Detail View Detail View Over View

Main View Main View

CommonTools

Left ViewTools

Right View Tools

Reference Sources

Reference (Ground) Coordinates File (Image) Coordinates


4. Ensure that image 11_188 is in the Left View and image 11_189 is in the Right View.

5. Adjust the brightness, then click Apply and review the modifications in the Viewer.

6. Click the Reset button to return the image to its original appearance.

If you had a large block of images, you would use these menus to 'scroll' through the images as you measure the GCPs.

Each image has its own brightness and contrast tools. This tool can be useful when viewing, identifying, and measuring GCPs.

At the bottom of the Point Measurement dialog, there are two CellArrays. Both are empty until points are input. The Reference CellArray (left), is where ground coordinates for each point are displayed.

Type Usage

Full - Control or check points with X, Y, and Z coordinates.

Control - Points used as control points during triangulation. (X, Y, Z are known.)

Horizontal - Points with X and Y coordinates. The Z coordinate is unknown and can be estimated during triangulation.

Check - Points are used to inde-pendently verify the quality of a tri-angulation. (X, Y, Z are known.)


The File CellArray, which is on the right, is where pixel coordinates are displayed. These coordinates are generated when you digitize the position of your Reference (Ground Control) points.

From what sources could we obtain reference coordinates?

What units are used in the File CellArray?

Vertical - Points with Z coordi-nates. The X and Y coordinates are unknown and can be estimated or ignored during triangulation.

Tie - Points appearing in the over-lapping areas of two or more images. (The X, Y, Z coordinates are unknown and determined dur-ing triangulation.)

None - Tie points with X, Y, and Z coordinates that are estimated dur-ing triangulation.

X Reference, Y Reference, Z Reference - Ground coordinates of each point.

Image Name - Name of the image where the ground control measurement was taken.

Active - X indicates that this measurement will be used in the aerial triangulation.

X File, Y File - Pixel coordinates (row and col-umn) of the measurement.


Task 2: Importing GCP Coordinates from an ASCII File

1. From the ERDAS menu , select View > View Text Files to open

the Text Editor.

2. Use the Open File icon and navigate to the LPS directory.

3. Select the denver_gcp.txt file, and click OK. The GCP ASCII file dis-plays.

4. Make note of the number of GCPs, then select File | Close in the Text Editor.

5. In the Common Tools section of the Point Measurement dialog, click the Import Points icon .


6. In the Import/Export Points dialog, click OK.

7. In the Reference ASCII File dialog, select denver_gcp.txt and click OK.

8. In the Reference Import Parameters dialog, click OK.

9. If an Attention dialog appears, click Yes.

10. In the Import Options dialog, click the View button. This will open the file in the Text Editor. Leave this Text Editor open.

11. From the text file, decide which is the appropriate Separator Character.

12. In the Import Options dialog, ensure that the Separator Character is set to WhiteSpace and that the Row Terminator Character is set to NewLine(Unix).

When would we need to edit the Field column?

13. Once the changes have been made, click OK in the Import Options. The selected columns in the Reference CellArray will populate.

14. Select File | Close in the Text Editor. The GCPs we input have X, Y, Z coordinates and will be used as control points.

In the Import Options dialog, listed under Column Mapping, are the selected columns from the Reference CellArray. Input Field Number is the denver_gcp.txt column number to be imported into the Reference CellArray.


What does the 'X' in the Active column mean? (Use On-Line Help)

Now that we have the ground coordinates of these points, you will measure them on the images.

15. Click the Viewing Properties icon .

16. In the Viewing Properties dialog, enable the Advanced radio button, and then enable the Residual checkbox. Click OK.

17. Click the Set Automatic (x,y) Drive icon . Following the second measured GCP, the position will automatically move to the next point.

A second way to import Reference points:

Click the Horizontal Reference icon .

From the GCP Reference Source dialog, enable the ASCII File (3D) radio button.

In the Reference ASCII File dialog, navigate to the LPS directory.

Select denver_gcp.txt and click OK.

In the Reference Import Parameters, click OK.

In the Import Options dialog, click OK.

The Point ID values would then need to be imported.


18. In the Reference (Ground) Coordinates section of the Point Mea-surement dialog, place the caret next to the first reference point.

19. Ensure that the Left View is set to 11_188.

20. In the overview for image 11_188, click and drag the cursor box to the first GCP location (visual aids to find the GCPs are provided at the end of this task for reference).

21. Use the Main View and Detail View windows to fine-tune the position-ing. If the location is difficult to see, adjust the contrast and/or bright-ness, and then click Apply to affect the image.

22. Click Reset to set both the contrast and brightness back to their origi-nal values.

23. Once you have located the point, click the Create Point icon , move the cursor to the point in the Detail View window.

24. Click to take a measurement. After the measurement is taken, a sym-bol and ID will display. Notice that the Image Name, along with X and Y File coordinates appear in the CellArray on the right.

25. Place the caret next to the second reference point. Use the previ-ous procedure to digitize the remaining GCP locations, generated from the ASCII file. When the GCP occurs in multiple images, digitize the point in the additional image(s).

26. When the points have all been measured, click the Save button.

If the caret is not moved to the first reference point, then the first measured GCP will be correlated to the incorrect refer-ence point and will need to be deleted and re-measured.

If the GCP needs to be moved, click to select the point then drag it to the correct location.

Remember to move the caret (>) to each point before digi-tizing.

Some GCPs will appear in all three images; you will need to use the Right View and Left View pull-down menus to view all the images.


X and Y Image Coordinates for the Ground Control Points

Point IDX File

11_188Y File

11_188X File

11_189Y File

11_189X File

11_190Y File

11_190

1 7300.869 1040.875 NA NA NA NA

2 7902.375 7432.905 NA NA NA NA

3 5876.865 5736.365 NA NA NA NA

4 3390.862 5161.596 6745.885 5107.455 NA NA

5 137.612 5240.366 3581.503 5119.324 6855.289 5224.150

6 NA NA 857.773 6577.114 4172.309 6653.906

7 NA NA 47.344 730.768 3558.664 885.625

8 NA NA NA NA 542.678 4238.856

9 NA NA NA NA 2358.746 7714.545


GCP Locations

GCP #1


GCP #2


GCP #3


GCP #4


GCP #5


GCP #6


GCP #7


GCP #8


GCP #9


Task 3: Adding a Reference image

All GCPs collected up to this point have been obtained from GPS coordinates. The next point you will collect will be from a reference image.

1. Click the Reset Horizontal Reference icon .

2. Ensure the Image Layer radio button is selected and click OK.

3. In the Reference Image Layer dialog, select park.img and click OK.

Under the Right View section, you will see that the Horizontal reference has been set to park.img. This image is a subset of an orthorectified IKONOS image. The area covered by this image can be found in all three input images.


4. From the Left View section, enable the Use Viewer As Reference checkbox.

The Left View has been changed to park. In the next few steps, you will obtain a GCP from the Reference image and then find that point in all of the input images. Currently, the input images are rotated at a 900 clockwise angle from park.img. In order to aid the locating of a common point, park.img will be rotated to match the input images.

5. RMB click in the Main View of park.img and select Rotate.

6. In the Rotate Image dialog, set the Rotation Angle to 90, enable the Clockwise radio button, and click the ApplyToLeft button.

7. Click Close.

8. Change the Right View to 11_188.


9. Click the Add button to add a tenth reference point.

10. Change the Type to Horizontal and the Usage to Control.

On the following page you will find the point to be digitized.

11. Click the Create Point icon and measure the GCP in the refer-ence image and in all of the input images.

12. When the point has been measured, click the Save button and then disable the Use Viewer As Reference checkbox.


Task 4: Adding Tie Points

Within the overlap between 2 or more images, you might be able to identify the same ground feature on different images. You do not need to know the coordi-nates (X,Y,Z) of this point. The Tie point is used during triangulation and X, Y, Z coordinates are generated for the Tie point when the triangulation is accepted.

You will use the same tools to add a Tie point to your set of reference points.

1. Using the Link Box, locate a feature in the two viewable images. Try and find a recognizable feature in both images using the image below as a guide.

2. From the Common Tools section, click the Add button again, then click the Create Point icon .


3. To keep this tool active, click the Keep Current Tool icon .

4. In the Detail View, place the cursor over the feature you identified and digitize the point. Immediately, move the cursor to the same fea-ture in the other image, and digitize again.

5. Repeat the above steps to digitize the Tie Point in the third image.

6. Click the Automatic Tie Properties icon and make sure that the following parameters are set:

7. Click OK in the Automatic Tie Point Generation Properties dialog.

The Type (None) and Usage (Tie) will default to the correct settings.

Tie points are very useful in the triangulation process, but they can be very time-consuming to collect. Leica Photo-grammetry Suite will automatically generate tie points. It uses a variation of Feature-Based Matching.

Auto Tie Point Collection - involves image matching as a means to identify ground features in overlapping areas of imagery. The various matching methods can be divided into three categories including:

Area based matching - Similarity of the gray level values within correlation windows based over the imagery.

Feature based matching - Point features are extracted and identified on adjacent images. The feature pair is given a correlation value. Points with correlation values, which are closer to 1, are recognized as better matches.

Relation based matching - Uses image features and the relationship between the features. Time-consuming but very accurate.

Images Used: All Available

Initial Type: Tie Points

Strategy Parameters: Avoid Shadow

Intended number of Points / Image:

25


8. Click the Perform Automatic Tie Generation icon . Tie Points will be generated on each of the images.

The diagram below shows the configuration required to perform automatic collec-tion on six overlapping images.

9. Click Close in the Auto Tie Summary dialog.

How many Tie points were generated?

Minimum Input Requirements for Automatic Tie Point Col-lection for a Frame Camera, digital camera, videography, or non-metric camera:

Initial approximations of ext.orientation (X, Y, Z, omega, phi, kappa), OR

At least two GCPs per stereo pair, OR

At least two (2) Tie Points per stereo pair


10. Visually inspect a sample of the collected tie points for accurate placement. If you find that the Tie Points are slightly off from the cor-rect position, adjust them.

11. Click the Select points common to both views icon . This will highlight all points that are found in the images currently displayed.

12. After all adjustments are completed, Save your measurements.

We have completed the ground point measurements. We are now ready to move on to the actual triangulation process, where a relationship is mathematically built between your coordinates.

If the Tie Points do not match, and there are gross errors, ask your instructor for assistance.


Class Notes


Exercise 7: Building the Relationship

Objective: To perform aerial triangulation in LPS, using fewer control points.

Task 1: Setup and Run Aerial Triangulation

You have now completed the ground point measurement and generated tie points. The block file is now ready to have the triangulation parameters set.

1. Ensure that the Point Measurement dialog is open, and the Refer-ence and File coordinates are visible.

2. Click the Triangulation Properties icon . The Aerial Triangula-tion dialog displays.

LPS uses a mathematical technique, known as Block Bun-dle Adjustment, for aerial triangulation. Block bundle adjust-ment performs three functions:

Determines the position and orienta-tion of each image at time of exposure

Determines coordinates for the tie points in overlapping areas

Identifies, removes, minimizes and distributes errors associated with the imagery, GCPs etc.

There should be Xs in the Active column beside all Control and Tie points. If there are not, click in the Active cells which do not have X's.


General Options

Maximum Iterations: Used in processing the aerial triangu-lation.

Convergence Value: The limit for the correction of ground coordinates in each iteration.

Point Options

Used to set the standard deviations (error estimates) asso-ciated with the image and ground coordinates.

Image Point Standard Deviation (pixels): During the triangu-lation process the image coordinate positions are allowed to fluctuate within the limits of these values. Larger values indicate poor image measurements.

Interior Options

Used to set the standard deviations for the interior orienta-tion parameters (millimeters).

Fixed for all Images: Self-calibration will not take place (i.e., the original focal length and principle point are maintained).

Exterior Options

Used to set the standard deviations for the exterior orienta-tion parameters, in meters and degrees.

Advanced Options

Additional Parameters Model: Used in the aerial triangula-tion for the compensation of systematic image errors.

Use Image Observations of Check Points in Triangulation: Can improve the precision of the solution.


3. In the General tab, leave the Maximum Iterations at 10 and the Convergence Value as 0.001.

How is the Convergence Value derived? (Use On-line Help.)

4. Click the Advanced Options tab and ensure that the Use Image Observations of Check Points in Triangulation checkbox is dis-abled.

5. Click Run to proceed with the triangulation process. When complete, a Triangulation Summary displays.

What is your Total Image Unit Weight RMSE?

During the triangulation process what is happening to the tie point reference coordinates?

6. Record the RMSE values in the space below. You will be comparing these values as you run through this exercise.

This form of Triangulation (Bundle Block Adjustment) uti-lizes an iterative least squares solution. The iterations will continue until the Maximum Iterations have been reached or the Convergence Value is less than 0.001.

As the iterations continue, the original coordinates are re-evaluated. If every difference between the re-evaluated and original values is less than the convergence value, the itera-tions will stop.

Total Image Unit-Weight RMSE: ______________


Following are two examples of Triangulation Summaries. Both RMSEs are expressed in pixels.

7. Look at the differences: While one has a good RMSE, the other has a poorer RMSE; however, both reached convergence.

8. Click the Review button.

This records the overall accuracy of each measured point.

Control Point RMSE: ________________________

________________________

________________________

________________________

________________________


9. Make a note of the points with a high Total RMSE (for both Ground Points and Image Points):

We will look at these in greater detail in the next exercise.

10. Close the Review Triangulation Results dialog.

11. Click Close in the Triangulation Summary dialog.

12. Click OK in the Aerial Triangulation dialog.

Task 2: Changing Triangulation Parameters

1. In the Point Measurement dialog, disable all of the Tie points and render them inactive by:

Highlight only the GCPs (i.e. Usage is set to Control) on the list by clicking the Point # with the Shift + LMB

RMB click on the highlighted Point # and select Invert Selection

Click on the Active column heading

RMB click over the Active column heading and select For-mula

Input a value of 0 in the Formula window and click Apply and Close

You will now see all of the selected rows, which are highlighted in yellow, become inactive.

2. Click the Perform Triangulation icon to re-run the aerial trian-gulation.

How have the triangulation results been modified?

3. Close the Triangulation Summary.

4. In the Point Measurement dialog, replace the 'X's beside the Tie points, making them active in the triangulation process. Use the pro-cess given above and add a 1 into the Formula text box.

__________________ __________________

__________________ __________________

__________________ __________________


5. Open the Triangulation Properties dialog and select the Point tab, then set the GCP Type to Same Weighted Values. This will allow the triangulation process to vary the GCP locations.

6. In the Aerial Triangulation dialog, click Run.

7. Close the Triangulation Summary, then click Cancel in the Aerial Triangulation dialog.

We have looked at the triangulation process, and some of the parameters that can be changed to improve the output results. We have not yet considered the possi-bility of poor GCPs.

The next exercise will examine the triangulation results in more detail, and search for poor GCP values.


Class Notes


Exercise 8: The Triangulation Results

Objective: Study the Triangulation Report results and make adjustments to improve the trian-gulation results.

Task 1: Analyze the Triangula-tion Report

The aerial triangulation report lists all of the input and output values used during the triangulation process. The report can be divided into several categories, but for a standard block the most significant are probably:

1. Click the Report of Triangulation Results icon . The Triangula-tion Report displays in the Text Editor.

2. In the Text Editor, select Find | Find, and in the edit box, type itera-tion, then click Find.

How many iterations did it take to reach the convergence value?

3. Look at the Standard Error for the last iteration. This value is impor-tant since it accumulates the effect of each image coordinate residual, to provide a global indicator of quality.

Control Point residuals

Check Point residuals

Image Coordinate results

By analyzing the Triangulation Report and finding the points with the most error (relatively large residuals) you can start to improve your triangulation.

Make note of the Output Units, mentioned at the top of the report.


What is the Standard Error following the last iteration and what are its units?

4. In the Search and Replace dialog, input rX in the top, then click Find.

5. Below, list the Point ID of the GCP with the largest residuals:

Point ID rX rY rZ

6. This point may fall on more than one image. To see residuals for each measurement, move to the residuals of image points section at the bottom of the Text Editor. Look for the Point ID listed above.

What Image were the poor residuals recorded on?

In what units are the residuals of image points?

7. Click Close in the Search and Replace dialog.

8. Select File | Close in the Text Editor.

Task 2: Removing GCPs and Re-run-ning the Triangula-tion

1. In the Reference CellArray of the Point Measurement dialog, click on the 'X' in the Active column for the row with the GCP mentioned above.

This point becomes inactive, and is no longer used in the triangulation.

During the triangulation process GCP values (X,Y,Z) are recalculated during each iteration. The newly calculated GCP coordinates are compared to the original reference GCP coordinates, and the differences (residuals) are listed.

This should take you to the portion of the report which con-tains the residuals of the control points.


2. Click the Triangulation icon . The Triangulation Summary dis-plays.

What is the new Total Image Unit-Weight RMSE?

3. Click the Report button. Use the scroll bar to move down through the triangulation report.

4. Take a look at the iteration and the residuals of the control points sec-tions.

Did it take fewer iterations to reach the convergence? How many?

5. Close the Text Editor.

Task 3: Accept and Update the Triangu-lation Results

In a production environment, you could go back and forth between removing, adjusting, or adding GCP's, and re-running the triangulation.

For this task, we will move on to accept the triangulation.

1. Click Update and Accept, then click Close in the Triangulation Summary.

What characteristics of the Tie points changed after Accepting?

2. Save then Close the Point Measurement dialog.

Why have the cells under Ext. changed from red to green?

3. From the LPS dialog, select File | Save.

LPS updated the X, Y, Z, Omega, Phi, and Kappa of each exposure, and the X, Y, and Z for the Tie Points. It also distributed the error around the images, thereby giving us better positional accuracy for the ortho-photo creation process.

Update to bring up-to-date the block file with the values cal-culated during the aerial triangulation. This will update the exterior orientation parameters (if set to Unknown or Initial), and interior orientation parameters (if they were estimated).

Accept to confirm the triangulation results and update the X, Y and Z tie point reference values.


Section 2: DTM Extraction

Section Objective

This section will examine the fundamental principles and science of extracting elevation data from stereo photographs.

Various tools and techniques will be introduced to evaluate the accuracy of the elevation data, and verify the resulting Digital Terrain Models (DTM).

Tools Used

DTM Extraction Tool

A component of LPS used to extract elevation mass points from a digital block.

Contour Generation

Used to create shapefile contours of the DTM.

Change Detection Tool

Tool used for comparing images for determining differences in pixel values.

96 Introduction to LPS Section 2: DTM Extraction


Exercise 1: In Search of Z

Objective: Focus on automatic DTM extraction, which is a capability added through LPS, and used prior to orthorectification.

Task 1: DEM Extraction Parameters

With the triangulation results accepted and the exterior orientation parameters updated, aerial triangulation is complete. The next step is often ortho-generation, which requires the use of a constant Z value, or the procurement of a DTM.

1. In the eWorkspace, click the ERDAS button and select Prefer-ences. The Preference Editor displays.

2. Change the Category to Viewer.

3. Scroll to Background Transparent and disable this option. This will ensure imagery displays properly in the Terrain Editor viewers.

4. Next, change the Category to LPS.

5. Scroll to DTM Minimum Overlap Percentage.

6. Change the DTM Minimum Overlap Percentage to 30.


7. Click User Save and Close.

8. From the LPS icon panel, click the DTM Extraction icon . The DTM Extraction dialog displays. You will be creating a DEM image file.

9. Click the Single DTM Mosaic radio button. This creates one output image.

10. For the Output DTM File name, type dem_20m_default.img.

11. Change the DTM Cell Size for X to a value of 20, and enable the Make Pixels Square checkbox. The setting for Units should remain as Meters.

12. Leave the DEM Background Value set at Default.

The cell size defaults to a value, which is approximately 10 times the spatial resolution of your image.

Generating exclusion areas in which DEM points would be generated, a default value would be assigned as the exclusion area elevation value.

Default Background Values will be 0 if all elevations are positive. If there are negative elevations, a DEM Back-ground value will be five (5) units [meters] lower than the minimum negative value.


13. Click the Run button to generate your DTM.

Following are the steps involved in the generation of a DEM.

Interest Point Determination - Each image in the block is processed to obtain a series of interest points. An interest point is a pixel which exhibits sufficient gray level variation when compared to a neighboring set of pixels.

After the LPS block has been triangulated, a triangle (epipo-lar plane) can be formed between two perspective centers (X, Y, and Z), and a ground feature. This plane is used to constrain the point search area on the adjacent image.

An interest point located on the reference image may have more than one possible match on the adjacent image. For each set of possible image points identified by LPS, a corre-lation coefficient is computed.

Epipolar Plane


Ground Point Determination - A Space Forward Intersection technique is used to compute the 3D coordinate associated with the mass points. This technique requires known exterior orientation parameters.

Collinearity Equations - Used to determine initial X Y Z ground point coordinates. The result is accurate ground coordinate X Y Z values for all ground points.

DTM Construction - When creating an output DEM, a temporary TIN is gener-ated, using the ground points from the above determination. A grid (X, Y spacing) user specified, is generated and interpolated over this temporary TIN.

A Z value (elevation value) is assigned to each grid point.

An output DEM image is generated.

Task 2: Viewing a DEM

You have created a default DEM. Later, you will compare the DEM to one that has been generated using the Advanced Properties.

1. In a Viewer, display your DEM, and start the Inquire Cursor.

2. Review the elevation values, then identify and make note of any anomalous areas that may have incorrect elevation values.

3. Close the Viewer.

Task 3: Advanced Options

The use of Advanced Options will more accurately, represent the Earth's surface when creating a DEM.

1. Repeat the steps of Task 1, but DO NOT DELETE the DEM you just created. Instead, type dem_20m_advanced.img as the name for the Output DTM File.

2. To open the DTM Extraction Properties dialog, click the Advanced Properties button.


The Advanced Properties are divided into five categories using the following tabs:

General

Image Pair

Area Selection

Accuracy

Seed Data

3. First, review the options in the General tab, and verify the projection. It should be the same as the projection of your LPS block.

4. The Horizontal Units and Vertical Units value should each be set to Meters.

5. Enable the Create Contour Map checkbox and use a Contour Inter-val of 5.00 (meters). The file created will be an ESRI 3D Shapefile.

6. Enable the Remove Contours Shorter Than checkbox, and input a value of 100. Any contour lines that are shorter than 100m will be removed.

By default, the interval is three times that of the Output DTM size.

By default, this is five times that of the image cell size.


7. Enable the Create DTM Point Status Output Image checkbox. This image is a visual representation of an accuracy report, which we will create in a later exercise. Based on correlation values obtained dur-ing Interest Point Matching, points will be color-coded:

Excellent = Light Green

Good = Medium Green

Fair = Yellow

Isolated = Orange

Suspicious = Red

8. Enable Reduce DTM Correlation Area by checkbox. Enter a value of 5%. .

9. Enable Trim DTM Border by checkbox. Enter a value of 5%. The DTM extraction area will be the full defined overlap, but will be reduced by 2.5% on all sides.

Task 4: Choosing Image Pairs

DEMs can be generated by using all or some Stereo Pairs.

1. Select the Image Pair tab. This CellArray shows all image pairs with an overlap of 50% or greater.

2. View your pairs by clicking the Open Viewer for Image Pair Selec-tion icon .

3. Click the caret to view the different Stereo Pairs.

4. Select both pairs by clicking in the Row column.

5. In the Image Detail column, RMB click a cell that has 100%. The options for this field display.

Correlation Area represents the output area of the DEM. A reduction will reduce the DEM area from the full stereo overlap.

Typically the outer edge of the stereo pair contains the larger errors. By this reduction, questionable areas and fiducial areas can be eliminated from the DEM..


6. For the Image Detail, select As Is to keep it at 100%. These values indicate the last pyramid layer to be used during the DTM mass point extraction.

On which pyramid layer does the correlation stop, if you input the value of 25% for Image Detail?

Task 5: Area Selec-tion and Identifica-tion

1. Select the Area Selection tab.

2. The Current Pair can be changed by using the drop down menu or the arrows. Currently the information relates to the first pair: 11_188_11_189

3. Click the Open the Viewer for Region Digitizing icon . A red outline displays, indicating the extent of your output DEM. This region has a Default search stragegy defined.

Since your image contains more developed urban areas, you will add another region with a different Region Strategy. For this pair, you will import an existing AOI.

4. Click the Load Regions from File icon , input denver-core.aoi, then click OK.

5. In the Area Selection CellArray, click in the Row column, and select Row 2, then RMB click on the Row number and choose Delete Selection.

6. In the Area Selection CellArray, click in the Row column, and select Row 2.

7. In the Region Strategy column, RMB click on the Default value and select High Urban.

8. Click the Edit Strategies icon . Compare and make note of the Search Size differences in both Default and High Urban.

9. Once finished examining the values, click the Cancel button to return to the DTM Extraction Properties window.

The DTM Correlation Area may alter the AOI shape.


10. Highlight Row 1 and set the Region Strategy to Low Urban.

The default Strategy parameters for a Low Urban can be modified.

11. Click the Edit Strategies icon . The Set Strategy Parameters dia-log displays.

12. Click the pull-down arrow for Strategy Name and select Low Urban, then review the Strategy parameters.

Is the Low Urban Search Size larger or smaller than the Default Search Size? Why?

13. Select the Low Urban strategy, and change the Search Size X value to 19.

Changes to the Correlation size and Search size are imple-mented by the strategy you choose.

Search Size - Window size is pixels searching along the epipolar line.

Correlation Size - A window, which moves inside the search area and is used to calculate a correlation coeffi-cient for interest points.

Correlation Coefficient - A threshold under which interest points are not used in LPS. The range is 0 to 1.

Topographic Type - Sets up internal parameters based on topography.

Object Type - DTM extraction algorithm will function more or less aggressively when selecting interest points. Flatter areas requires fewer interest points than mountainous areas.

Allow Adaptive Change - LPS can make changes to the various parameters as the DTM is computed. This may improve the results of the strategy application. Adaptive changes take place between iterative pyramid layer pro-cessing.

Use Image Band - Select the layer you want the correlation to use.

DTM Filtering - High - Filters out elevation spikes and pits. Low - Quicker, but possibly leaves more elevation spikes.


14. Click the pull-down arrow for DTM Filtering and select Moderate.

15. Click OK. When prompted whether to modify Strategy Parameters, select Yes.

16. RMB click in the Row column and choose Select None.

17. Click the Goto Next Pair arrow (down) to display the second stereo pair. As you did previously, set the Default Region Strategy to Low Urban.

You will now digitize the second strategy area in this pair.

18. In the Overview window, move the cursor to the bottom-right part of your image and resize the box, if needed.

19. Click the Create Polygon Regions icon and digitize around the downtown core area in the Viewer. The image will scroll as the mouse is moved outside of the Viewer.

20. Double-click to close the polygon.

21. Click to select/highlight the polygon. If editing is required, click the Reshape icon and edit as needed.

22. Change the Region Strategy of your new polygon to High Urban.


23. Move the caret to Row 2, and then click the Edit Strategies icon .

24. Change the Strategy Name to High Urban.

25. Change the Search Size X to 33, change the Filtering to Moderate, and then click OK.

26. When prompted whether to modify Strategy Parameters, select Yes.

Task 6: The Region Z

The next few steps will step through how to use the Region Z column.

1. Digitize another polygon. It doesn't have to represent anything, since you will soon delete it.

2. In the Exclude Area row, click on the word Undefined, then select Custom.

For what could the Region Z Value dialog be used?

3. Click OK in the Region Z Value dialog.

4. Highlight this polygon, and then use the Cut Selected Regions icon to delete it.

5. Click OK in the DTM Extraction Properties dialog.

By default, the Region Strategy is set to Exclude Area.

Custom is only available if the Region Strategy is set to Exclude Area.


6. Click Run in the DTM Extraction dialog to create your second DEM.

Why was the X Search Size increased in the Low Urban Strategy parameters?

7. When the DTM Extraction is finished, go to the LPS menu and select File | Save then select File | Close.

The Accuracy options will be examined in a subsequent exercise.


Exercise 2: Visualizing Your DEM

Objective: LPS tools will be used to compare DEMs and identify areas with questionable ele-vation values.

Task 1: Evaluate Using a Contour File

Two DTMs have been generated. The first, dem_20m_default.img, has only default properties. The second, dem_20m_advanced.img, has modified, advanced properties.

1. In separate Viewers, open each DTM, and list below, any differences you see.

In your opinion, which image looks more reliable?

2. In the Viewer containing dem_20m_advanced.img, open the shape-file, dem_20m_advanced.shp. (You will need to change the Files of type to Shapefile.)

This is a 3D shapefile, containing X, Y, and Z values. The contour may be difficult to see, so you will modify the colors.

3. Make sure that dem_20m_advanced.shp is the topmost layer in the Contents pane. Then, in the Symbology group on the Drawing tab,

select Vector Symbology , then select Automatic | Unique Val-ues. This allows colors to be added, based on unique attributes.

4. In the Unique Value dialog, ensure that the Unique Value is set to Z, then click OK.

_______________________________________________________

_______________________________________________________

_______________________________________________________


What is the range of contour values?

5. RMB click in the Row column and click Select All to select all the rows.

6. RMB click on any of the Symbol cells, then click Other.

7. Change the Outer Color to Green and the Outer Width to 1, click Apply, then click OK.

8. From the Symbology dialog, click Apply.

You will now change the color of individual contours.

9. Select the rows containing Z values ending in 00, 25, 50, and 75, then RMB click in the Row column and select Invert Selection.

10. RMB click in one of the selected cells, and from the pull-down list, change the color to Solid Red.

11. In the Symbology dialog, click Apply then Close. Select Yes to save the changes and name the Symbology: contour.evs

How might the contours help to identify potential problem areas?

Task 2: Evaluate Using a Relief Image

1. Open two more 2D Views. Select one Viewer, and from the Quick Access menu, click the Open Layer icon .

2. Select dem_20m_default.img, then click the Raster Options tab.

3. Click the pull-down arrow for Display As, select Relief, then click OK.

4. In the other 2D View, click the Open Layer icon , select dem_20m_advanced.img, and use the same Relief option as above, to display the image.

In comparing the two relief images, are there differences you had not yet noticed when viewing the DEM images earlier?


Task 3: Evaluate Using Change Detec-tion Tools

Until now, most of the identifiable change has been in the downtown area, as it was more difficult to identify change in the flatter terrain. To help better identify change, the Change Detection tool is used.

1. From the eWorkspace, select the Raster tab, and select Change

Detection > Zonal Change > Image Difference.

2. Input dem_20m_default.img as the Before Image, and dem_20m_advanced.img as the After Image.

3. For the Image Difference file, type dem_diff.img, and for the High-light Change file, type dem_high.img.

4. Change the Increase and Decrease values from the default to a value of 1%. This will highlight within the images, any el

Date post:	30-Sep-2015
Category:	Documents
Upload:	luis-felipe-pinzon-uribe
View:	344 times
Download:	13 times

Manual de ERDAS

Documents