regiongrow算法matlab

Region growing算法是一种基于像素相似度的图像分割算法。在MATLAB中，可以使用以下代码来实现Region Growing算法：

function [segmented_img, count] = regiongrow(img, seeds, thresh)
% img: 输入的图像矩阵
% seeds: 种子点的坐标，格式为[n,2]，n为种子点的数量
% thresh: 阈值

% 初始化
[row, col] = size(img);
segmented_img = zeros(row, col);
count = 0;

% 计算种子点的灰度值
seed_gray_vals = zeros(size(seeds, 1), 1);
for i = 1:size(seeds, 1)
    seed_gray_vals(i) = img(seeds(i,1), seeds(i,2));
end

% 遍历种子点
for i = 1:size(seeds, 1)
    seed_row = seeds(i,1);
    seed_col = seeds(i,2);
    
    % 初始化队列
    queue = [seed_row, seed_col];
    
    % 遍历队列
    while size(queue, 1) > 0
        % 出队列，获取当前像素点坐标
        current_pixel = queue(1,:);
        queue(1,:) = [];
        row = current_pixel(1);
        col = current_pixel(2);
        
        % 判断当前像素点是否已经分割
        if segmented_img(row, col) == 0
            % 计算当前像素点与种子点的差值
            diff = abs(img(row, col) - seed_gray_vals(i));
            
            % 判断差值是否小于阈值
            if diff <= thresh
                % 记录分割后的像素点
                segmented_img(row, col) = i;
                count = count + 1;
                
                % 将当前像素点的邻居加入队列
                if row > 1
                    queue = [queue; row-1, col];
                end
                if row < size(img,1)
                    queue = [queue; row+1, col];
                end
                if col > 1
                    queue = [queue; row, col-1];
                end
                if col < size(img,2)
                    queue = [queue; row, col+1];
                end
            end
        end
    end
end

在使用时，可以调用该函数并传入图像矩阵、种子点坐标和阈值。函数会返回分割后的图像矩阵和分割区域的数量。