matlab实现灰度图像的最大熵阈值分割

灰度图像的最大熵阈值分割是一种常用的图像分割方法，它可以将图像分成两类，以使得分割后的图像信息熵最大。以下是 MATLAB 实现最大熵阈值分割的基本步骤：

1. 读取灰度图像并将其转换为二维数组。

img = imread('image.jpg');
gray_img = rgb2gray(img);
[m, n] = size(gray_img);

2. 计算图像直方图并归一化。

histogram = imhist(gray_img);
normalized_hist = histogram / (m * n);

3. 初始化最大熵和最大熵阈值。

max_entropy = 0;
max_threshold = 0;

4. 遍历所有可能的阈值，并计算分割后的两个子图像的信息熵。

for threshold = 1:255
    % 计算两个子图像的像素个数和像素值出现次数
    pixel_count = zeros(2, 256);
    for i = 1:m
        for j = 1:n
            pixel_value = gray_img(i, j);
            if pixel_value <= threshold
                pixel_count(1, pixel_value + 1) = pixel_count(1, pixel_value + 1) + 1;
            else
                pixel_count(2, pixel_value + 1) = pixel_count(2, pixel_value + 1) + 1;
            end
        end
    end
    % 归一化两个子图像的像素值出现次数
    normalized_pixel_count = pixel_count / (m * n);
    % 计算两个子图像的信息熵
    entropy = -sum(sum(normalized_pixel_count .* log2(normalized_pixel_count + eps)));
    % 更新最大熵和最大熵阈值
    if entropy > max_entropy
        max_entropy = entropy;
        max_threshold = threshold;
    end
end

5. 根据最大熵阈值对图像进行分割。

binary_img = gray_img > max_threshold;

完整的代码如下：

img = imread('image.jpg');
gray_img = rgb2gray(img);
[m, n] = size(gray_img);

histogram = imhist(gray_img);
normalized_hist = histogram / (m * n);

max_entropy = 0;
max_threshold = 0;

for threshold = 1:255
    pixel_count = zeros(2, 256);
    for i = 1:m
        for j = 1:n
            pixel_value = gray_img(i, j);
            if pixel_value <= threshold
                pixel_count(1, pixel_value + 1) = pixel_count(1, pixel_value + 1) + 1;
            else
                pixel_count(2, pixel_value + 1) = pixel_count(2, pixel_value + 1) + 1;
            end
        end
    end
    normalized_pixel_count = pixel_count / (m * n);
    entropy = -sum(sum(normalized_pixel_count .* log2(normalized_pixel_count + eps)));
    if entropy > max_entropy
        max_entropy = entropy;
        max_threshold = threshold;
    end
end

binary_img = gray_img > max_threshold;

figure;
subplot(1, 2, 1);
imshow(gray_img);
title('原图');
subplot(1, 2, 2);
imshow(binary_img);
title('分割后的二值图');

运行结果如下图所示：