视频图像差分检测

 

简介

视频图像序列处理方法中，差分算法以其高效性、简便性应用广泛。这里给出应用实例，并给出实验结果。

 

代码

% By lyqmath

clc; clear all; close all;
% 原始视频
targetavi = 'traffic.avi';
% 检测结果视频
resultavi = 'c:\\result.avi';
%% 读取视频
mov = mmreader(targetavi);
fnum = mov.NumberOfFrames;
%% 建立结果视频
aviobj = avifile(resultavi);
aviobj.Quality = 100;
aviobj.Fps = 25;
aviobj.compression = 'Indeo5';
%% 帧间差分法
figure(1);
for i = 2 : fnum
    x = read(mov, i-1);
    y = read(mov, i);
    subplot(1, 2, 1); imshow(x, []); title(sprintf('第%d帧视频，By lyqmath', i-1), 'FontWeight', 'Bold', 'Color', 'r');
    % 灰度化
    if ndims(x) == 3
        m = rgb2gray(x);
    else
        m = x;
    end
    if ndims(y) == 3
        n = rgb2gray(y);
    else
        n = y;
    end
    % 中值滤波
    m = medfilt2(m);
    n = medfilt2(n);
    % 数据类型转换
    q = im2double(n);
    w = im2double(m);
    % 差分
    c = q-w;
    % 阈值，此值可以调节
    t = 40/256;
    % 阈值分割
    c(abs(c)>=t)=255;
    c(abs(c)<t) = 0;
    c = logical(c);
    x1 = x(:, :, 1); x2 = x(:, :, 2); x3 = x(:, :, 3);
    x1(c) = 255; x2(c) = 0; x3(c) = 0;
    xc = cat(3, x1, x2, x3);
    subplot(1, 2, 2); imshow(xc, []); title(sprintf('第%d帧视频识别结果，By lyqmath', i-1), 'FontWeight', 'Bold', 'Color', 'r');
    f = getframe(gcf);
    f = frame2im(f);
    % 生成视频
    aviobj = addframe(aviobj, f);
end
%% 关闭视频句柄
aviobj = close(aviobj);

结果

http://blog.sina.com.cn/lyqmath