Identify colour of each pixel using iterative statement

Hi Everyone,

Absolute begginer here.

I am looking to write an iterative statement (loop function) to identify pixels that are 'red', and change them to 'green' in a new image. Pixels that are not identified as 'red' need to be left alone. I understand this is not the best way to do this, but I need to explore the limitations of this process before moving to more advance image processing tools. Keep in mind that I don't have impixel() function as I don't have the image processing toolbox yet. I have a RGB 512 x 512 image.

My theshold would be that the pixel value is:

red > 180 && green < 50 (I'm not concerned with blue at the moment)

Logically the statement would read as follows:

If pixel red value > 180 and pixel green value < 50, then multiply red value by 0.5 and green value by 2.0.

The part I am having trouble with is actually building this into the code and creating a function that goes through pixel by pixel and creates a new image were the 'red' pixels are changed to 'green'

Thank you all in advance.

What I have so far is below:

% % % Code to import image and modify colour

% File name to call
filename = 'pepper.bmp';

% Function to call filename
OriginalImage = imread(filename);

    % image(OriginalImage) % Displays Original Image (Discovered 'image'
    % fucntion does not display as clear of a picture as 'imshow'

% Extract individual RGB colour channels
redChannel = OriginalImage(:, :, 1);
greenChannel = OriginalImage(:, :, 2);
blueChannel = OriginalImage(:, :, 3);


% Construct new image by analysing pixel by pixel
   for
      ???? 
   end

% Display both images side by side
subplot(1,2,1), imshow(OriginalImage)
title('Original Image');
subplot(1,2,2), imshow(ModifiedImage)
title('Modfied Image');

Answer 1

[Ameer Hamza]

In MATLAB, you don't need to write for loops to do pixel manipulation. Following code shows one of the way

im = imread('peacock.jpg');

red = im(:,:,1);
green = im(:,:,2);
blue = im(:,:,3);

mask = (red > 180) & (green < 50);

red(mask) = red(mask)*0.5;
green(mask) = green(mask)*2;

im_new(:,:,1) = red;
im_new(:,:,2) = green;
im_new(:,:,3) = blue;

[darova]

>> 2 > 1 * 3
ans =
     0

[Image Analyst]

You can stay using logicals and must cast back to the original class (assuming uint8):

mask = (red > 180) & (green < 50);
red(mask) = uint8(red(mask)*0.5);
green(mask) = uint8(green(mask)*2);

[Ameer Hamza]

Thanks for pointing out. I was actually trying different things and accidently pasted the above version instead of and (&) version.

Answer 2

[Kyle Robertson]

Thank you so so much to everyone that contributed! This worked great! The code that I ended up with is below:

I do know that iterative statements were not nessesary but that was part of the excersie.

Best Wishes!!

% % % Code to import image and modify colour

% File name to call
filename = 'pepper.bmp';
OriginalImage = imread(filename);  %stored as a 512x512x3 matrix array  % 3 dimensional array

ModifiedImage = OriginalImage;

%Condition Statement for which pixels to process
for i = 1:512                %iterate over each row
    for j = 1:512            %iterate over each column
        if OriginalImage(i,j,1)>OriginalImage(i,j,2)*1.15             
            
            ModifiedImage(i,j,1) = (OriginalImage(i,j,2)*0.90)+60; %Swap Red and Green Channels and apply linear scaling
            ModifiedImage(i,j,2) = (OriginalImage(i,j,1)*1.08)-50;     
            ModifiedImage(i,j,3) = (OriginalImage(i,j,3)*0.90)+50; %Scale the Blue Channel for more accurate appearance
            %Scale Brightness of processed area. Also useful for debugging selection area by setting to 0
            ModifiedImage(i,j,:) = (ModifiedImage(i,j,:)*1.2)-30;                
        end
   end
end

% Display both images side by side
figure('Name','Peppers','WindowState','Maximized')
subplot(1,2,1), imshow(OriginalImage)
title('Original Image');
subplot(1,2,2), imshow(ModifiedImage)
title('Modfied Image');