How can I create a luminosity histogram of an image without the green background?

This is an image like the ones I will be trying to apply this with. The section I'm trying to make the histogram for is always going to be grayscale.

This is what I tried but it creates a histogram that is obviously wrong:

rgbImage2 = imread('Screenshot 2022-01-18 150342._SubCu1#01-25-2022 16-08.bmp');

redChannel2 = rgbImage2(:, :, 1);

greenChannel2 = rgbImage2(:, :, 2);

blueChannel2 = rgbImage2(:, :, 3);

binaryRed = redChannel2 == 0;

binaryGreen = greenChannel2 == 255;

mask = binaryRed & binaryGreen;

isnan(mask);

histogram(binaryImage(~mask), 50)

%histogramArea = ~mask;

%hist(histogramArea)

Would appreciate any help, Thanks!

Answer 1

[Image Analyst]

Try this:

rgbImage = imread('image.bmp');
subplot(2, 2, 1);
imshow(rgbImage);
impixelinfo;
title('Original RGB Image')
% Get green mask
[r,g,b] = imsplit(rgbImage);
mask = r == 0 & g == 255 & b == 0;
subplot(2, 2, 2);
imshow(mask);
impixelinfo;
title('Green Mask Image')
% Get "luminosity image"
labImage = rgb2lab(rgbImage);
luminosityImage = labImage(:, :, 1);
subplot(2, 2, 3);
imshow(luminosityImage, []);
impixelinfo;
title('Luminosity Image')
% Get the histogram where the mask is false (the hole in the green mask);
edges = 0 : 256;
pixelCounts = histcounts(luminosityImage(~mask), edges);
subplot(2, 2, 4);
bar(edges(1:end-1), pixelCounts);
grid on;
title('Luminosity Histogram of Masked Image')

[Justin Luk]

Thank you for your help so far! This is definitely a lot closer than I was before

The issue I seem to have now is that it is not possible to obtain values larger than 100 on the histogram even when a completely white image is used. Not sure why that is happening. Would it be correct to just multiply all values by 2.55?

[Image Analyst]

I used your exact image you attached so I don't know why it didn't work. Did you use some function on it like double(), im2double(), mat2gray(), rescale() or anything like that?

[Justin Luk]

When I ran that code with the image I attached, it looked the same way as it did in your screenshot.

The last screenshot I attached was me testing to see if the luminosity histogram was accurate by plugging a purely white image into the program to see if it would generate a histogram that had only values at 255. However, it generated a histogram only at a value of 100 which shouldn't be correct.

Thanks

[Image Analyst]

Why do you think that a completely white image maxes out at 255 should not have a "luminosity" of 100%? So you think it should be a lesser percentage? If so, why? What definition of luminosity are you using? Actually luminosity has units of watts and is the total amount of power emitted by an object producing electromagnetic radiation.

https://en.wikipedia.org/wiki/Luminosity

[Justin Luk]

Ah I see, I didn’t realize it was percentage rather than the actual brightness value. That makes sense, thanks

[Image Analyst]

If you'd rather use

% Get "luminosity image"
luminosityImage = rgb2gray(labImage);

instead of

% Get "luminosity image"
labImage = rgb2lab(rgbImage);
luminosityImage = labImage(:, :, 1);

then you can do that. It gives you values in gray levels according to a weighted sum of the separate color channels, instead of a "percent reflectance" like rgb2lab() gives.

[Justin Luk]

Got it, thank you!