MATLAB实现改进Otsu算法的代码

MATLAB实现改进Otsu算法的代码

1. 参数设置

% 读取图像
I = imread('example_image.jpg'); % 替换为实际图像路径
I = rgb2gray(I); % 转换为灰度图像
I = im2double(I); % 转换为双精度浮点数

2. 计算图像均值

% 计算图像均值
meanIntensity = mean(I(:));

3. 改进的Otsu算法

% 改进的Otsu算法
function threshold = improvedOtsu(I, meanIntensity)
    % 获取图像大小
    [rows, cols] = size(I);
    
    % 初始化变量
    totalPixels = rows * cols;
    maxVariance = 0;
    threshold = 0;
    
    % 计算直方图
    hist = imhist(I);
    cumulativeHist = cumsum(hist);
    
    % 计算全局均值
    totalMean = sum((0:255) .* hist) / totalPixels;
    
    % 在均值到最大灰度值之间搜索最佳阈值
    for T = meanIntensity:255
        % 计算前景和背景的像素数量
        backgroundPixels = cumulativeHist(T);
        foregroundPixels = totalPixels - backgroundPixels;
        
        % 避免除以零
        if backgroundPixels == 0 || foregroundPixels == 0
            continue;
        end
        
        % 计算前景和背景的均值
        backgroundMean = sum((0:T-1) .* hist(1:T)) / backgroundPixels;
        foregroundMean = sum((T:255) .* hist(T+1:end)) / foregroundPixels;
        
        % 计算类间方差
        variance = backgroundPixels * foregroundPixels * (backgroundMean - foregroundMean)^2;
        
        % 更新最大方差和阈值
        if variance > maxVariance
            maxVariance = variance;
            threshold = T;
        end
    end
end

% 调用改进的Otsu算法
threshold = improvedOtsu(I, meanIntensity);

4. 图像分割

% 使用阈值进行图像分割
segmentedImage = I > threshold;

5. 显示结果

% 显示原始图像和分割后的图像
figure;
subplot(1, 2, 1);
imshow(I);
title('Original Image');

subplot(1, 2, 2);
imshow(segmentedImage);
title('Segmented Image');

参考代码 ostu图像分割阈值算法