How to enhance an image with respect to an image of much better quality.

Question

SCOTT LIM 2017-3-9

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/328912-how-to-enhance-an-image-with-respect-to-an-image-of-much-better-quality

评论： Scott Lim 2017-3-9

Hi, i'm working on a project to enhance a 16 bit greyscale image. These images come from a scanning machine that produces a header file with the relevant Min/Max contrast pixel values.

I have scanned the same item on 2 different machines to find microdefects. However, one of the machines doesn't perform well. Is there a way to enhance the defects of the poorer image to something similar that of the better quality image?

From the picture attached, Original 03 is the image from good machine. Original 04 is the image from the bad machine.

I have tried using imhistmatch() but it doesn't enhance the bad image enough. Furthermore, in order to use imhistmatch, i have to first convert both images into

Im03 = mat2gray(Im03,[MinCon3 MaxCon3]);
Im04 = mat2gray(Im04,[MinCon4 MaxCon4]);
imhistmatch(Im04,Im03,65535);

Also, I have tried enhancing the contrast, whites and blacks and offset gains. Including using filters such as wiener2, guassfilt2 and median. They do not benefit the imhistmatch() results.

Thank you.