One problem was the lack of good comments. But the main problem was not casting your logical (binary) images to uint8 before creating the output RGB image. Here is the fixed code:
% Read in the input RGB image.
inputRGBImage = imread('peppers.png');
% Extract separate color channels.
rmat = inputRGBImage(:,:,1);
gmat = inputRGBImage(:,:,2);
bmat = inputRGBImage(:,:,3);
% Threshold each color channel to select pixels darker than 50 gray levels.
% Then set those pixels to 0 (black) in the original color channels.
binaryImageR = rmat < 50;
rmat(binaryImageR) = 0;
binaryImageG = gmat < 50;
gmat(binaryImageG) = 0;
binaryImageB = bmat < 50;
bmat(binaryImageB) = 0;
% RGB reconstruction section - create sum images but ignore them and don't use them.
% Find pixels that are darker than 50 in all 3 color channels, but ignore this binary image.
imgSum1 = (binaryImageR & binaryImageG & binaryImageB);
% Now do something weird, but never use this image.
imgSum2 = (rmat & gmat & bmat); % ?????
% Now create an RGB image showing the binary masks for each color channel.
rgbImage = cat(3, uint8(255 * binaryImageR), uint8(255 * binaryImageG), uint8(255 * binaryImageB)); % Make the RGB image after filter
% Display images.
figure;
subplot(2,3,1), imshow(rmat);
title('Red Plane');
subplot(2,3,2), imshow(gmat);
title('Green Plane');
subplot(2,3,3), imshow(bmat);
title('Blue Plane');
subplot(2,3,4), imshow(inputRGBImage);
title('Original RGB Image');
subplot(2,3,5), imshow(rgbImage);
title('Binary Images');
