I want to code an image stabilizer but when I try to copy an RGB image to a m*n*3 zeros matrix and display it I get a weird result.

My problem is in the "% Corrected image" section, here is my code (I attached the images I'm using):

close all; clear all; clc;

% Image loading

imds = imageDatastore('blablabla');

images = readall(imds);

% Centroid calculation

centroid1 = centroidFinder(images{1});

centroid2 = centroidFinder(images{2});

% Centroid difference

centroidD = centroid2-centroid1;

% Corrected image

sizeI = size(images{1});

correctedDim1 = sizeI(1)+abs(centroidD(1)); correctedDim2 = sizeI(2)+abs(centroidD(2)); correctedDim3 = sizeI(3);

correctedI = zeros(correctedDim1, correctedDim2, correctedDim3);

correctedI(1:1:sizeI(1), 1:1:sizeI(2), :) = images{2};

imshow(correctedI);

function centroid = centroidFinder(image)

%{ NOTE: Add a range for G2 %}
 
[R,G,B] = imsplit(image); % RGB image split
 
G2 = G == max(max(G)); % Discard of elements not belinging to G

C = regionprops(true(size(G2)),G2,'WeightedCentroid'); % Centroid calculation
 
centroid = [round(C.WeightedCentroid(1)), round(C.WeightedCentroid(2))]; % Centroid vector output
 
end

Answer 1

[DGM]

Since the expected scale of image data depends on the class of the array, you need to make sure that the array is the appropriate class for the data you're putting in it.

correctedI = zeros(correctedDim1, correctedDim2, correctedDim3, 'uint8');

[Reinhold Erwin Suchowitzky Toba]

Thank you! It works