matlab Otsu法代码实现

以下是 Matlab 中 Otsu 算法的实现代码：

function [threshold, varargout] = otsu(I)
% Otsu's thresholding algorithm
% Usage: [threshold] = otsu(I)
%        [threshold, varargout] = otsu(I)
% Inputs:
%   I           -   input image (grayscale or RGB)
%
% Outputs:
%   threshold   -   Otsu's threshold
%   varargout   -   additional outputs (optional)
%                   - histogram: normalized histogram of the image
%                   - omega: probability mass function
%                   - mu: cumulative mean
%
% Author: Shanmuganathan Raman
% Date: October 20, 2021

% Check if input is grayscale
if size(I,3) == 3
    I = rgb2gray(I);
end

% Calculate histogram
h = imhist(I);
h = h/sum(h);

% Initialization
omega = zeros(256,1);
mu = zeros(256,1);

% Calculate cumulative probability
omega(1) = h(1);
mu(1) = 0;
for i = 2:256
    omega(i) = omega(i-1) + h(i);
    mu(i) = mu(i-1) + (i-1)*h(i);
end

% Calculate variance for all possible thresholds
sigma_b_squared = zeros(256,1);
for i = 1:256
    if omega(i) ~= 0 && omega(i) ~= 1
        sigma_b_squared(i) = ((mu(end)*omega(i) - mu(i))^2) / (omega(i)*(1-omega(i)));
    end
end

% Find threshold with maximum variance
[~, threshold] = max(sigma_b_squared);

% Output additional variables if requested
if nargout > 1
    varargout{1} = h;
    varargout{2} = omega;
    varargout{3} = mu;
end

该函数接受一个灰度图像作为输入，并返回 Otsu 阈值。如果需要，还可以返回其他变量，例如图像的直方图、概率质量函数和累积均值。

使用示例：

I = imread('cameraman.tif');
threshold = otsu(I);

在这个例子中，我们使用 `imread` 函数加载一个名为 `cameraman.tif` 的图像，并将其传递给 `otsu` 函数。函数返回图像的 Otsu 阈值。