How to detect circle in this pic?

I want detect every circle in this pic using this code. But, I detect only four.

so, How to detect every circle in this pic? what's wrong my code?

clear

clc

image = imread('example5.PNG');

image = edge(imgaussfilt(rgb2gray(image), 2), 'Canny');

[x y] = find(image == 1);

pindex = cat(2, y, x);

hough_matrix = zeros(1000, 1000);

rad = [55 56 64 65 51 52];

pindex = pindex + 245;

for k = 1 : 7

for p=1:size(pindex, 1)

center = pindex(p, :);

% 모든 허프 matrix에 대한 이중 for문

for i = center(1)-rad(k):center(1)+rad(k)

x = i - center(1);

for j = center(2)-rad(k):center(2)+rad(k)

y = j - center(2);

if round(sqrt(x*x + y*y)) == rad(k)

hough_matrix(j, i) = hough_matrix(j, i) + 1;

end

end

end

end

[x y] = find(hough_matrix == max(hough_matrix, [], 'all'));

center = cat(2, y, x);

center = center - 245;

if k==1

figure;imshow(image);

else

viscircles(center, rad(k));

end

end

Answer 1

[Shashank Gupta]

Hi Kim,

Looks like you are trying to implement a Hough transform algorithm from the scratch. MATLAB already has the function “imfindcircles” which detect the circle in the image given a range of radius. Also this function uses the same Hough transform algorithm. But before using the function you need to do some pre-processing, I am following up your code to find the edge, now use this edged image to fill up the holes and now you can use “imfindcircle” function. This might help you.

I am also attaching a snippet of the code for the reference

image = imread('image.png');
image = edge(imgaussfilt(rgb2gray(image), 2), 'Canny');

img = imfill(image,'holes');
imshow(img);