How to determine the aspect ratio of the detected components and throw the components which does not meet the threshold

I have a code attached to which does the bounding box on rgb image. First I convert yellow portion to white and rest all to black. then mask it to find the bounding box. How to determine the aspect ratio of the detected components(white image) and throw the components which does not meet the threshold.

Answer 1

[Madhav Thakker]

Hi

I’m assuming you want to extract the bounding box from the binarized image and then calculate the aspect ratio of the detected components for further processing.

After getting the binarized image from yellow.m, you can use imdilate with a structuring element as ones(19) to merge all the white components. After that you can just use regionprops to get the region of the detected component. I am also attaching the code for easy understanding.

Hope this helps.

I = imread('fogg1.jpg'); 
J = rgb2ycbcr(I); 
CB = J(:,:,2); 
histCB = imhist(CB)/numel(CB); 
CCB = zeros(size(histCB)); 
for j = 1:length(CCB) 
    CCB(j) = sum(histCB(1:j)); 
end 
inx = find(CCB < 0.01,1,'last'); 
B = CB <= inx-1; 
  
  
se = ones(19); 
dilated = imdilate(B, se); 
info = regionprops(dilated,'Boundingbox') ; 
imshow(dilated) 
hold on 
for k = 1 : length(info) 
     BB = info(k).BoundingBox; 
     rectangle('Position', [BB(1),BB(2),BB(3),BB(4)],'EdgeColor','r','LineWidth',2) ; 
end 
  
lower_left_corner = [BB(1), BB(2)]; 
width = BB(3); 
height = BB(4); 

For more information, refer to imdilate MATLAB documentation - https://in.mathworks.com/help/images/ref/imdilate.html