Contrast Enhancement Transformation for Remote Sensing Images

Project Discription

Abstract

As with the help of internet transfer of data is getting easy. So the number of internet users are increasing day by day. As different images from different sources like satellite, MRI reports, etc have some hidden information which need to be improved for better understanding. Therefore enhancement of image plays important role in these area. So contrast enhancement feature is necessary to improve through image processing. As contrast feature increase the visibility of the image and small region of the image which are not visible in original is highlighted and overall quality of poor images get improve. So enhancement of images of different users for various resources is selected for research in this work.  Proposed work uses Logarithmic function for contrast image enhancement. Here image undergoes logarithmic transformation then diffusion of image is done into three parts for adjusting the contrast finally after contrast adjustment fusion is done. Experiment is done on standard as well as artificial image sets. Results shows that proposed work better on different evaluation parameters as compare to previous algorithms as this improves the visualization of the image for showing some of the dull information which is obtain by making some transformation.

Logarithmic Transformation:

In this transformation of image pixel values is done. As this transformation can directly modify the pixel value so some precaution is taken during this process. Chance of new color of the pixel is totally depend on the previous value, so an average value of the surrounding pixels is taken for uniformity of the transformation. This can be understand as let I image has pixel position (m, n) then its transformation is done by below steps.

Image De-Fusion

Here image is divided into three part base on the new transform values. As new pixel value is compare for three categories first is low pixel value second is middle and third is high intensity value. Base on the two threshold value one can classify image or defused.

Adjust Contrast

Now contrast of the image is adjusted by equation 4.2 and 4.3. For the global contrast enhancement, the knee transfer function stretches the low-intensity range by determining knee points according to the dominant brightness of each layer. More specifically, in the low-intensity layer, a single knee point is computed as

Pl = bl+ wl*(ml-bl)

Matlab Projects With Source Code

Project Sample Image