Contrast stretching??? this pic

Answer 1

[DGM]

Let's start with the basic approach using stretchlim() and imadjust(). In this case, stretchlim() returns the limits that will remap the extrema of the source to black and white -- or more accurately, it will remap the intensity values such that the bottom and top 1% of pixels are clipped. This is done on all three channels independently; consequently, there is a noticeable color shift, and the brightest and darkest regions shift toward pure white and black.

inpict = imread('https://www.mathworks.com/matlabcentral/answers/uploaded_files/388723/image.jpeg');

limits = stretchlim(inpict,0.01);
outpict = imadjust(inpict,limits);

imshow([inpict; outpict])

Alternatively, one can try to reduce the independent limits returned by stretchlim() to a single set of limits that can be applied uniformly to all three channels. In this case, I just chose to use the mean, but one could also use the inner or outer extrema. Note that there isn't as much of a hue shift, though there might be a slight shift in saturation.

inpict = imread('https://www.mathworks.com/matlabcentral/answers/uploaded_files/388723/image.jpeg');

limits = mean(stretchlim(inpict,0.01),2);
outpict = imadjust(inpict,limits);

imshow([inpict; outpict])

Of course, one doesn't have to do this in RGB. In this example, L is stretched in LAB.

inpict = imread('https://www.mathworks.com/matlabcentral/answers/uploaded_files/388723/image.jpeg');

outpict = rgb2lab(inpict);
limits = stretchlim(outpict(:,:,1)/100,0.01);
outpict(:,:,1) = imadjust(outpict(:,:,1)/100,limits)*100;
outpict = im2uint8(lab2rgb(outpict));

imshow([inpict; outpict])

There's nothing saying that you have to adhere to those limits either. You can adjust the input endpoints to whatever you want:

inpict = imread('https://www.mathworks.com/matlabcentral/answers/uploaded_files/388723/image.jpeg');

limits = [0.2; 0.8];
outpict = imadjust(inpict,limits);

imshow([inpict; outpict])