can't understand the cause of this error, trying to crop multi blobs in one image

hello, i'm working with video motion, here it can be the case where we found many object mobile, i used blobAnalysis to delimited them, i calcul for each one it variance, i used this code but an error appeared, can anyone help me please:

foregroundDetector = vision.ForegroundDetector(*....);
videoReader = vision.VideoFileReader('*.avi',...
    'VideoOutputDataType','uint8');
J=0;
i=0;
while ~isDone(videoReader);
    J=J+1
    frameRGB = step(videoReader); % read the next video frame
    foreground = step(foregroundDetector, frameRGB);

blobAnalysis = vision.BlobAnalysis('BoundingBoxOutputPort', true, ...
    'AreaOutputPort', false, 'CentroidOutputPort', false, ...
    'MinimumBlobArea', 1700,'MaximumCount',2);
bbox = step(blobAnalysis, foreground);
result = insertShape(frameRGB, 'Rectangle', bbox, 'Color', 'green');
num = size(bbox,1);

    if J>=1;

      if num>1
      for i=1: num
      croppedImage = frameRGB(bbox(i, 2):bbox(i, 2)+bbox(i, 4), bbox(i, 1):bbox(i, 1)+bbox(i, 3), :);
      blobVariance(i) = var(double(croppedImage(:)))

    end
result = insertShape(frameRGB, 'Rectangle', bbox, 'Color', 'red');
result = insertText(result, [10 10], numperso, 'BoxOpacity', 1, ...
    'FontSize', 14);
figure(1); imshow(result);

blobVariance=[];
croppedImage=[];

    end

end
  end

end and the error is:

Index exceeds matrix dimensions.
Error in MaxvariancecorriG (line *)
 croppedImage = frameRGB(bbox(i, 2):bbox(i, 2)+bbox(i, 4), bbox(i, 1):bbox(i, 1)+bbox(i, 3),
 :);

[Stefan Raab]

Hey bay rem,

I am not familiar with the "step(videoPlayer)" function, but I suppose it results the frame as a matrix? Your error says, that you try to address an index which does not exist. Does "step" return a two dimensional array? Then you have too much indices in the "croppedImage = ..." command.

Kind regards, Stefan

Answer 1

[Walter Roberson]

   croppedImage = frameRGB(bbox(i, 2):bbox(i, 2)+bbox(i, 4)-1, bbox(i, 1):bbox(i, 1)+bbox(i, 3)-1,:);

Remember, a width of 1 would mean that you should end at the same pixel, so your endpoint would be 0 further on from where you are. A width of 2 would mean you would end at the next pixel, so your endpoint would be 1 further on from where you are. And so on: a width of N requires a last index of N-1 beyond where you are.