【含操作视频】基于harris角点的2张图像配准算法matlab仿真

function [corners, im] = extractHarrisCorner(img, thresh) dx = [-1 0 1; -1 0 1; -1 0 1]; % Derivative masks dy = dx'; Ix = conv2(img, dx, 'same'); % Image derivatives Iy = conv2(img, dy, 'same'); Ix2 = Ix .^ 2; % Products of derivatives at every pixel Iy2 = Iy .^ 2; Ixy = Ix .* Iy; Sx2 = conv2(Ix2, ones(3), 'same'); %The sums of the products of derivatives at each pixel (3 * 3 neighbourhood) Sy2 = conv2(Iy2, ones(3), 'same'); Sxy = conv2(Ixy, ones(3), 'same'); numOfRows = size(img, 1); numOfColumns = size(img, 2); % im = zeros(numOfRows, numOfColumns); for x=1:numOfRows for y=1:numOfColumns H = [Sx2(x, y) Sxy(x, y); Sxy(x, y) Sy2(x, y)]; % Define the matrix Harris at each pixel K = det(H)/trace(H); % The Harris response measure im(x, y) = K; end end % output = im > imdilate(im, [1 1 1; 1 0 1; 1 1 1]); radius = 3; % Size of mask. mx = ordfilt2(im, radius^2, ones(radius)); % Grey-scale dilate. corner = (im == mx) & (im > thresh); % Find maxima (with non-maximum suppresion). % corner = im > thresh; % without non-maximum suppresion [r,c] = find(corner); % Find row,column coords. corners = [r, c]; corners(find(corners(:,1)<5),:) = []; % Remove corners which are on edge of images. corners(find(corners(:,2)<5),:) = []; corners(end-3:end,:) = []; % Removed a few indices that have something wrong end