超像素分割slic的matlab代码

以下是SLIC超像素分割的MATLAB代码：

%% SLIC超像素分割

clc; clear all; close all;

% 读取图像
img = imread('lena.bmp');
figure; imshow(img); title('原始图像');

% 设置参数
num_superpixels = 1000; % 超像素数量
compactness = 10; % 超像素紧密度，越大则超像素更规则

% 计算步长
[h, w, ~] = size(img);
step = sqrt(h*w/num_superpixels);

% 初始化超像素分割结果
labels = zeros(h, w);

% 初始化超像素中心
centers = step/2:step:w;
centers = repmat(centers, [ceil(h/step), 1]);
centers = centers(1:h, :);

% 迭代优化
for i = 1:10
    % 计算超像素中心所在的网格位置
    gridx = floor(centers(:)/step)+1;
    gridy = floor((1:h)'/step)+1;
    
    % 扩展图像边界
    img_ext = padarray(img, [step, step], 'symmetric', 'both');
    gridx_ext = padarray(gridx, [step, step], 'symmetric', 'both');
    gridy_ext = padarray(gridy, [step, step], 'symmetric', 'both');
    labels_ext = padarray(labels, [step, step], 'symmetric', 'both');
    
    % 计算每个超像素中心附近的像素点
    for j = 1:num_superpixels
        % 确定超像素中心的位置
        cx = centers(j);
        cy = find(gridy_ext(:, cx) == j, 1, 'first');
        cy = cy - step;
        
        % 计算超像素中心周围的像素点
        x1 = max(cx-step, 1);
        x2 = min(cx+step, w)+step;
        y1 = max(cy-step, 1);
        y2 = min(cy+step, h)+step;
        pixels = img_ext(y1:y2, x1:x2, :);
        labels_pixels = labels_ext(y1:y2, x1:x2);
        [yy, xx] = find(labels_pixels == j);
        pixels = pixels(min(yy):max(yy), min(xx):max(xx), :);
        labels_pixels = labels_pixels(min(yy):max(yy), min(xx):max(xx));
        
        % 计算每个像素点与超像素中心的距离
        [h_p, w_p, ~] = size(pixels);
        dists = zeros(h_p, w_p);
        for k = 1:h_p
            for l = 1:w_p
                dists(k, l) = sqrt((k-yy(1))^2 + (l-xx(1))^2) + sqrt((pixels(k, l, 1)-pixels(yy(1), xx(1), 1))^2 + (pixels(k, l, 2)-pixels(yy(1), xx(1), 2))^2 + (pixels(k, l, 3)-pixels(yy(1), xx(1), 3))^2)/compactness;
            end
        end
        
        % 更新像素点的标签
        labels_pixels_new = labels_pixels;
        [~, ind] = sort(dists(:));
        ind = ind(1:numel(yy));
        for k = 1:numel(yy)
            [y, x] = ind2sub([h_p, w_p], ind(k));
            labels_pixels_new(yy(k), xx(k)) = labels_pixels(y, x);
        end
        
        % 更新超像素标签
        labels_ext(y1:y2, x1:x2) = labels_pixels_new;
    end
    
    % 缩小图像边界
    labels = labels_ext(step+1:h+step, step+1:w+step);
    
    % 更新超像素中心
    for j = 1:num_superpixels
        [yy, xx] = find(labels == j);
        centers(j, :) = [mean(xx), mean(yy)];
    end
end

% 显示超像素分割结果
figure; imshow(labels, []); title('超像素分割结果');

该代码实现了SLIC超像素分割算法，包括计算超像素中心、计算每个超像素周围的像素点、计算像素点与超像素中心的距离、更新像素点的标签和更新超像素中心等步骤。其中，使用了MATLAB自带的`padarray`函数对图像边界进行了扩展和缩小操作。