Detecting Corners using CornerMetric

The cornermetric function identifies corners in an image. The two algorithms employed by the function—the Harris corner detection method (the default) and Shi and Tomasi's minimum eigenvalue method—both use the eigenvalues of the summation of the squared difference matrix (SSD). The eigenvalues of an SSD matrix represent the differences between the surroundings of a pixel and the surroundings of its neighbors. The larger the difference between the surroundings of a pixel and those of its neighbors, the larger the eigenvalues. The larger the eigenvalues, the more likely a pixel is at a corner.


I = imread('1.jpg');
C = cornermetric(I, 'SensitivityFactor', 0.01);
imshow(I);
[y, x] = find(corner_peaks);
hold on
plot(x, y, '.', 'Color', 'y')
hold off