How to regroup or reconstruct connected elements in a binary image into squares
2 views (last 30 days)
Show older comments
Hello, please I need help with this.
I have a binary image(50x50).
The matrix was randomly generated therefore making the image noisy. My aim is to regroup or reconstruct connected elements of same value together into different square shapes(not necessarily a perfect square), the edges of the square doesn't have to be smooth since the matrix is randomly generated. I did the following but I want to reconstruct all same elements into squares.
I = imread('Fig1.jpg');
BW = imbinarize(I); % Convert image to binary matrix.
L = bwlabeln(BW, 26) % Label connected elements.
4 Comments
KALYAN ACHARJYA
on 30 Aug 2021
After seeing the image, I request you to rethink the question? What is your actual objective?
Answers (1)
Image Analyst
on 30 Aug 2021
Why are you using bwlabeln()? For one thing, that's for 3-D images, not 2-D like you have. And what do you mean "of the same value"? Originally, you only have two values 0 and 1. Do you simply want a square that is all white with as many white dots there are in your original image?
numDots = nnz(BW);
rows = ceil(sqrt(numDots));
whiteSquare = ones(rows, rows);
numBlack = rows * rows - numDots;
if numBlack >= 1
whiteSquare(1:numBlack) = 0;
end
Of course if you label them each connected component will have it's own ID number. Not sure if you want 4 connected or 8 connected but I believe if you use 26 connected on a 2-D image you'll get 8 connected (if it doesn't throw an error).
12 Comments
Image Analyst
on 1 Sep 2021
Don't use the while. It's not needed and confusing since h<0 is a matrix -- a lot of values -- not a single value. What Walter gave you is sufficient and correct.
h(h == -1) = 0;
Again, delete the while and end lines.
See Also
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!