otsu阈值分割 matlab 代码

Otsu阈值分割是一种自适应的图像二值化方法，它通过最大化类间方差来确定一个最佳的阈值，将图像分为背景和前景两个部分。

以下是使用Matlab实现Otsu阈值分割的代码：

function [threshold, segmented_img] = otsu_thresholding(img)
    % 转换为灰度图像
    gray_img = rgb2gray(img);
    
    % 计算灰度直方图
    hist = imhist(gray_img);
    
    % 归一化直方图
    hist_norm = hist / numel(gray_img);
    
    % 初始化类内方差和总方差
    within_class_var = zeros(256, 1);
    total_var = sum((0:255).^2 .* hist_norm) - sum((0:255) .* hist_norm).^2;
    
    % 遍历所有可能的阈值，计算类内方差
    for threshold = 0:255
        % 计算类内方差
        w0 = sum(hist_norm(1:threshold+1));
        w1 = 1 - w0;
        u0 = sum((0:threshold) .* hist_norm(1:threshold+1)) / w0;
        u1 = sum((threshold+1:255) .* hist_norm(threshold+2:256)) / w1;
        within_class_var(threshold+1) = w0 * w1 * (u0 - u1)^2;
    end
    
    % 找到使类间方差最大的阈值
    [~, threshold] = max(within_class_var);
    threshold = threshold - 1; % 索引从0开始
    
    % 根据阈值进行分割
    segmented_img = imbinarize(gray_img, threshold/255);
end

使用方法：

img = imread('your_image.jpg');
[threshold, segmented_img] = otsu_thresholding(img);
imshow(segmented_img);

希望以上代码能对您有所帮助！