How to find local minima in a image histogram?

Hello, I have few search about it in google. but i'm confused :( I would like to search for the first (leftmost) local minimum in the image histogram. For example the attached image is histogram of sample image. and we need local minima. I need just one number. Thanks

Answer 1

[Image Analyst]

Use diff() to get the difference between the n+1st bin and the nth bin. Then use find to get the index

[counts, grayLevels] = imhist(grayImage);
index = find(diff(counts) >= 0, 1, 'first');
minGL = grayLevels(index);

[Walter Roberson]

Should that be

 index = find(diff(counts) >= 0, 1, 'first');

[Image Analyst]

Yes, thanks for correcting. I'll fix it.

[Dani D]

Thanks, But just 0 or 1 is result. :( For example i test cameraman.tif and minGL is 0

[Walter Roberson]

That would not be uncommon. Suppose an image had no true black, then the first histogram entry would be a count of 0. The very first shade that did occur would have a non-zero entry, which is an increase. You would have a local minimum of 0 from the first bin to the first bin that had some data, at which point you would have an increase.

[Image Analyst]

Yes, that's the first place where the histogram is flat. Perhaps you want findpeaks() where you get more control over how big a peak or valley you detect.