Hi,
starts by reading the image and converting it to grayscale if it is not already. To enhance the visibility of features, adjusts the image contrast and applies Canny edge detection. Circle detection is then performed in two passes with different radius ranges ([15 30] and [30 50]) to capture circles of varying sizes. The detected circles from both passes are combined and sorted by their strength metric to prioritize the most confidently detected circles. Finally, displays the original image and overlays the strongest detected circles, ensuring that a specified number of circles are shown, thereby improving the overall detection accuracy for all the coins in the image.
% Read the image
A = imread('coins.png');
% Check if the image is already grayscale
if size(A, 3) == 3
% Convert to grayscale if it is an RGB image
grayA = im2gray(A);
else
% Use the image as is if it's already grayscale
grayA = A;
end
% Adjust contrast
adjustedA = imadjust(grayA);
% Apply edge detection
edgesA = edge(adjustedA, 'canny');
% First pass of circle detection
[centers1, radii1, metric1] = imfindcircles(edgesA, [15 30], 'Sensitivity', 0.9);
% Second pass of circle detection with a different range
[centers2, radii2, metric2] = imfindcircles(edgesA, [30 50], 'Sensitivity', 0.9);
% Combine results from both passes
centers = [centers1; centers2];
radii = [radii1; radii2];
metric = [metric1; metric2];
% Sort circles based on metric
[~, sortedIndices] = sort(metric, 'descend');
centers = centers(sortedIndices, :);
radii = radii(sortedIndices);
metric = metric(sortedIndices);
% Check if enough circles are found
numCircles = min(10, length(radii)); % Adjust the number of circles to display as needed
% Display the original image
imshow(A);
hold on;
% Display the strongest circles
if numCircles > 0
centersStrong = centers(1:numCircles, :);
radiiStrong = radii(1:numCircles);
metricStrong = metric(1:numCircles);
viscircles(centersStrong, radiiStrong, 'EdgeColor', 'b');
else
disp('No circles detected.');
end
hold off;
