混合高斯背景减除法matlab代码

以下是混合高斯背景减除法的 MATLAB 代码：

clear;

% 读入视频文件
v = VideoReader('test.mp4');

% 初始化混合高斯模型参数
K = 5; % 高斯分布数量
alpha = 0.01; % 学习率
T = 0.5; % 阈值
w = ones(1, K) / K; % 权重
mu = zeros(K, 3); % 均值
sigma = zeros(K, 3); % 方差

% 读入第一帧并初始化混合高斯模型参数
frame = readFrame(v);
[h, w, ~] = size(frame);
bg = double(frame);
fg = zeros(h, w);
for i = 1:h
    for j = 1:w
        for k = 1:3
            mu(1, k) = mu(1, k) + bg(i, j, k);
            sigma(1, k) = sigma(1, k) + bg(i, j, k)^2;
        end
    end
end
mu = mu / (h * w);
sigma = sigma / (h * w) - mu.^2;

% 处理每一帧并进行背景差分
while hasFrame(v)
    frame = readFrame(v);
    diff = zeros(h, w, K);
    for i = 1:K
        diff(:, :, i) = abs(double(frame) - mu(i, :));
    end
    [~, index] = min(diff, [], 3);
    mask = max(abs(double(frame) - mu(index, :)), [], 3) > T;
    fg = mask;
    for i = 1:h
        for j = 1:w
            if mask(i, j) == 1
                for k = 1:3
                    mu(index(i, j), k) = (1 - alpha) * mu(index(i, j), k) + alpha * double(frame(i, j, k));
                    sigma(index(i, j), k) = (1 - alpha) * sigma(index(i, j), k) + alpha * (double(frame(i, j, k))^2);
                    w(index(i, j)) = (1 - alpha) * w(index(i, j)) + alpha;
                end
            else
                w(index(i, j)) = (1 - alpha) * w(index(i, j));
            end
            w = w / sum(w);
            mu(index(i, j), :) = w(index(i, j)) * double(frame(i, j, :)) + (1 - w(index(i, j))) * mu(index(i, j), :);
            sigma(index(i, j), :) = w(index(i, j)) * (double(frame(i, j, :)).^2) + (1 - w(index(i, j))) * sigma(index(i, j), :) - mu(index(i, j), :).^2;
        end
    end
    bg = uint8(mu);
    imshow(fg);
end

其中，`test.mp4` 是需要进行背景差分的视频文件名，`K` 是高斯分布数量，`alpha` 是学习率，`T` 是阈值。该代码会对视频文件进行背景差分并输出前景掩码。