image intensity scale sub-divide into two ranges

I have an image. I obtain the histogram for that image and calculate probability distribution using the code below:

if size(I,3) == 1 %grayscale image
    [n_countR, x_valueR] = imhist(I(:,:,1));
Nt = size(I,1) * size(I,2); %Total number of pixels in the image ROW X COL
%Lmax segmenting color levels 0-256
Lmax = 256;   %256 different maximum levels are considered in an image (i.e., 0 to 255)

% Probability distribution for each intensity level histogram 0 - 256 
for i = 1:Lmax
    if size(I,3) == 1 
        %grayscale image
        probR(i) = n_countR(i) / Nt;
end

Now I want this image histogram, having range 0 to 255, to sub-divide into two having the range 0 to 127 and 128 to 255 respectively and find probability distribution for both. Please help. I hope now you can better understand my question. Much thanks..

Answer 1

[Guillaume]

By default, imhist on a greyscale image outputs 256 bins, just as you want. You then jut have to split this output in half, using simple indexing.

if size(I, 3) == 1  %in that case I(:,:, 1) == I
  h = imhist(I); %histogram over 256 bins by default
  hlow = h(1:128); %lower part of the histogram
  hhigh = h(129:256); %higher part of the histogram
end

To convert to pdf over the total number of pixels, use matrix operations rather than for loops:

pdflow = hlow/numel(I);
pdfhigh = hhigh/numel(I);

or if you want pdfs over the population of each half-histogram:

pdflow = hlow/sum(hlow);
pdfhigh = hhigh/sum(hhigh);

Answer 2

[Thorsten]

h1 = hist(I(I<=127), 0:127);
h2 = hist(I(I>=128), 128:255);

[Nataliya]

the histogram for h1 ranges between 0 to 1. isn't it show the range between 0 to 127?

[Thorsten]

Make sure that your image I is in the range 0:255. For example

 I = imread('peppers.png');

Don't use im2double which converts the image to the range 0:1.

Alternatively, if your image is in the range 0:1, use

I = I*255;

before you run my code.

[Nataliya]

I have edit my question. Kindly review it and answer. I will be thankful to you

[Thorsten]

 ind = find(I < 128); 
 Plow = hist(I(I < 128), 0:127)/numel(ind);
 ind = find(I >= 128); 
 Plow = hist(I(I >= 128), 128:255)/numel(ind);