HIGH DYNAMIC RANGE OF MULTISPECTRAL ACQUISITION … · Selvam Indian Research Publications @ Selvam...

International Journal of Innovative Research in Engineering Science and Technology

APRIL 2018 ISSN 2320 –981X

Selvam Indian Research Publications @ Selvam Educational Institutions IJIREST Vol VI Issue 02 PP 18-30

HIGH DYNAMIC RANGE OF MULTISPECTRAL ACQUISITION USING SPATIAL IMAGES

1M.Kavitha , M.Tech., 2N.Kannan, M.E., and 3S.Dharanya, M.E.,

1Assistant Professor/ CSE,

Dhirajlal Gandhi College of Technology,

Salem-636309, India.

[email protected]

2Assistant Professor/ ECE,

Mahendra Institute of Technology,

Mallasamuthram, Namakkal- 637 503,

India.

[email protected]

3Assistant Professor/ CSE,

Dhirajlal Gandhi College of Technology,

Salem-636309, India.

[email protected]

*Corresponding Author

e-mail: [email protected]

Contact: +91-7339303819

mailto:[email protected]




APRIL 2018 ISSN 2320 –981X


ABSTRACT

This paper introduces an effective technique to enhance the spatial images. Multiple exposure of

PAN images are collected in the broad visual wavelength range but rendered in gray scale images.

During this process, displacements of the images caused by object movements often yield motion blur

and ghosting artifacts. The resultant output is low resolution values. To address the problem, this paper

presents an efficient and accurate multiple colored image fusion technique to bringing out the high

dynamic range of images. The captured different views of spatial images are multiplied by pixel based

multiplication techniques. Wavelet fusion method and morphological reconstruction brings high

resolution image.

Keyword:

PAN images,

Pixel based multiplication,

Wavelet fusion,

Morphological reconstruction,

Erosion


APRIL 2018 ISSN 2320 –981X


I. INTRODUCTION

In different angle of any viewing condition, the human visual system can capture a wide dynamic

range of irradiance (about 14 orders in log unit), whereas the active range of charge-coupled device or

matching semiconductor sensors in most of today’s cameras does not cover the perceptional range of

real scenes. It is important in many applications to capture a wide range of irradiance of natural scene

and store it as a pixel. In the application of CG, a high dynamic range image is widely used for high-

quality rendering (display) with image-based lighting.

Nowadays, HDR imaging technologies have been developed and some sensors are commercially

available. They are used for in-vehicle cameras, surveillance in night vision, camera-guided aircraft

docking, high-contrast photo development, robot vision, etc. In the last decade, to capture the HDRI,

many techniques have been anticipated based on the multiple-exposure principle, in which the HDRI is

constructed by merging some photographs shot with multiple exposures. Many of the techniques assume

that a scene is static during taking photographs. The motion of objects causes motion blur and ghosting

artifacts. Although in some fields, such as video coding and stereo vision, many displacement (or

motion) estimation methods are proposed; simply applying them into the multiple exposure fusion often

fails since the intensity levels of the images are significantly different due to the failure of camera

response curve estimation, and more importantly, low and high exposure causes blackout and whiteout

to some regions of the images, respectively, in which correspondence between the images is hard to

find. Moreover, in the case of low exposure, noises such as thermal noise and dark current sometimes

make the displacement estimation difficult. None of the conventional methods addresses all of the

problems. In this paper, we propose an algorithm of the HDRI estimation based on the Markov random

field model.

We can construct the HDRI by taking into consideration displacements, underexposure and

overexposure (saturation), and occlusions. The displacement vectors, as well as the occlusion and the

saturation, are detected by the MAP estimation. In our method, we do not need to estimate accurate

motion vectors but displacement to the pixel with the closest irradiance, whereas the conventional

methods such as try to accurately estimate the motion. This relaxation improves the final quality of the

HDRI. The occlusion and the saturation are clearly classified and then separately treated, which results

in the accurate removal of ghosting artifacts. In the following section, we introduce a technique for


APRIL 2018 ISSN 2320 –981X


combining the multiple exposure images. We point out that weighting functions used in the conventional

methods have a drawback in a case of capturing a scene with movement and then propose a new

weighting function. A pixel based multiplication and morphological erosion technique are proposed in

Section V and VI. In Section VII we show some experimental results to confirm the validity of our work

and then, we conclude our work in section VIII.

II. ALGORITHM

Step 1: Preprocessing

Step 2: De-noising

Step 3: Pixel Based Multiplication

Step 4: Morphological Erosion

Step 5: Wavelet Fusion

Preprocessed

images

filtered

images

multiplied RGB merged

images images

Eroded images

INPUT

IMAGE

DENOISING

TECHNIQUE

PREPROCESSING

TECHNIQUE

PIXEL BASED

MULTIPLICA-

TION

TECHNIQUE

RGB

CONVERSION

IMAGES

OUTPUT

IMAGES

MORPHOLOGICAL

TECHNIQUE

Figure.1. Architecture Diagram

III. PREPROCESSING

Preprocessing helps for the improvement of the image data that suppresses unwanted distortions

or enhances some image features important for further processing.

There are two steps in preprocessing,

Acquisition

Spatial images are usually large in its memory, before using those images; it has to be reduced by

the compression method.


APRIL 2018 ISSN 2320 –981X


Image Registration

It is used in medical and satellite imagery to align images from different camera sources. It helps

overcome issues such as image rotation, scale, and skew that is common when overlaying

images.

Figure.2. Preprocessing

IV. DENOISING

It is a process of removing noise from the spatial image. There are two effective techniques to remove

salt and pepper noise in the image.

Median filter

Median filter is a noise removal technique which removes salt and pepper noise without reducing

the image sharpness. The median filter considers each pixel in the image in turn and looks at its nearby

neighbors to decide whether or not it is representative of its surroundings. Instead of simply replacing

the pixel value with the mean of neighboring pixel values, it replaces it with the median of those values.

The median is calculated by first sorting all the pixel values from the surrounding neighborhood

into numerical order and then replacing the pixel being considered with the middle pixel value. (If the

neighborhood under consideration contains an even number of pixels, the average of the two middle

pixel values is used).

Gaussian filter

Gaussian filter is probability density function equal to that of the normal distribution over the image. A

special case is white Gaussian noise, in which the values at any pair of times are identically distributed

and statistically independent (and hence uncorrelated). In communication channel testing and modeling,

Gaussian noise is used as additive white noise to generate additive white Gaussian noise


APRIL 2018 ISSN 2320 –981X


Figure.3. De-noising

V. PIXEL BASED MULTIPLICATION

Pixel based multiplication image is arithmetic operators, multiplication comes in two main forms.

The first form takes two input images and produce an output images in which the pixel value are just

those of the first image, multiplied by the values of the corresponding values in the second images. The

second form takes a single input image and produce output in which each pixel values is multiplied by a

specified constant. This latter form is probably the more widely used and is generally called scaling.


APRIL 2018 ISSN 2320 –981X


Figure.4. Pixel based multiplication

Steps:

Preprocessed image are converted into R image, G image and B image.

The two images individual R image, G image and B image are multiplied through the algorithm

pixel based multiplication. Pixels of RGB images are multiplied of the enhancement of the spatial

image.

Multiplied R image, G image and B images are combined together using image fusion method. This

kind of fusion provides clear and detailed pixel values of spatial images.

VI .MORPHOLOGICAL EROSION

In the erosion process, the image has been shrink or it removes the boundaries of the images

which will sharpen the resultant image. The number of pixels removed from the objects in an image

depends on the size and shape of the structuring element used to process the image.

Figure.5. Morphological Erosion

The erosion of A by B expression:

Where,

A is the fused image,

B is a structuring element,

Disk shape structuring element is used for erosion with the fused image.

Through this process the resultant image get sharpened in its nature. It will give the clear crystal

clear spatial image as output.

AƟB= ∩b⊂B Ab


APRIL 2018 ISSN 2320 –981X


VII. EXPERIMENTAL RESULT

(a) (b)

Figure. 6. (a) Front View Image-Preprocessed Image (b) Side View Image - Preprocessed Image

(c)


APRIL 2018 ISSN 2320 –981X


Figure.. 7.(C) Salt and pepper noise removed image

(d)

Figure.. 8. (d) Gaussian and Median filter

(e)

Figure.. 9. (e) Histogram of the noise removed image


APRIL 2018 ISSN 2320 –981X


(f) (g)


APRIL 2018 ISSN 2320 –981X


(h) (i)

Figure.10.(f) RGB Conversion-Red Channel (g) RGB Conversion-Green Channel

(h) RGB Conversion-Blue Channel (i) Eroded Image


APRIL 2018 ISSN 2320 –981X


(j)

Figure..12.(j). Resultant Image-Enhanced Spatial Image

VIII.CONCLUSION

The project entitled high dynamic range of multispectral acquisition using spatial images is done

in effective manner. This project will be highly user friendly and makes the users to select the images to

be fused and the performance of various algorithms can be valued by the human perception. The fusion

methods used in this proposed system is pixel based multiplication, morphological reconstruction. The

images are captured using RGB conversion images and then applied to the pixel based multiplication by

using a wavelet transformation to gives a fused images. Then the resultant image is applied to the

morphological Erosion. Because of the benefits of image fusion although higher and higher resolution

images obtained in the output. Aiming at the limitations of existing fusion methods, this paper proposes

a new fusion method which combines pixel based multiplication and morphological operation.

The future work can be enhanced with the technique called dilation using different algorithm or

can use dictionary training model, where the clustering of the source images can be performed and

trained with Orthogonal matching pursuit or FOCUSS algorithm.


APRIL 2018 ISSN 2320 –981X


IX. REFERENCES

Barata T, and Pina P, Sep (2013), ‘Morphological approach for feature space partitioning’, IEEE

Geosci. Remote Sens. Lett., vol. 3, no. 1, pp. 173–177.

Bin Yang and Shutao Li, Member, IEEE , april .(2010) ‘Multifocus Image Fusion and

Restoration With Sparse Representation’ IEEE transactions on instrumentation and

measurement, vol. 59, no. 4.

Naidu V.P.S September (2011), ‘Image Fusion Technique using Multi-resolution singular Value

Decomposition’, Defence Science Journal, pp. 479-484, vol. 61, no. 5.

Nannan yu, Tianshuang qiu, Feng bi, and Aiqi wang, September. (2011) ‘image features

extraction and fusion based on joint sparse representation’ ieee journal of selected topics in

signal processing, vol. 5, no. 5.

Prakash N.K July (2011), ‘International Journal of Enterprise Computing and Business Systems’,

ISSN, vol. 1 issue 2.

Sagar BSD, Gandhi G, and Rao BSP (2012), ‘Applications of mathematical morphology on

water body studies’, Int. J. Remote Sens., vol. 16, no. 8, pp. 1495–1502.

List of Figures

1. Figure.1. Architecture Diagram

2. Figure.2. Preprocessing

3. Figure.3. De-noising

4. Figure.4. Pixel based multiplication

5. Figure.5. Morphological Erosion

6. Figure.6. (a) Front View Image-Preprocessed Image (b) Side View Image - Preprocessed

Image

7. Figure.7.(C) Salt and pepper noise removed image

8. Figure.8. (d) Gaussian and Median filter

9. Figure.9. (e) Histogram of the noise removed image

10. Figure.10.(f) RGB Conversion-Red Channel (g) RGB Conversion-Green Channel (h) RGB

Conversion-Blue Channel (i) Eroded Image

11. Figure.12.(j). Resultant Image-Enhanced Spatial Image

Date post:	29-Sep-2020
Category:	Documents
Upload:	others
View:	4 times
Download:	0 times

HIGH DYNAMIC RANGE OF MULTISPECTRAL ACQUISITION … · Selvam Indian Research Publications @ Selvam...

Documents