基于视觉工具箱和背景差法的行人检测,行走轨迹跟踪,人员行走习惯统计matlab仿真

1.算法运行效果图预览

(完整程序运行后无水印)

在三维图中，幅度越大，则表示人员更习惯的行走路线。

2.算法运行软件版本

matlab2022a

3.部分核心程序

（完整版代码包含详细中文注释和操作步骤视频）

    for j=1:length(inds)
        % 调整边界框格式
        [R_,C_] = find(lb == inds(j));  
        upLPoint  = min([R_ C_]);  
        dWindow   = max([R_ C_]) - upLPoint + 1;  
        box       = [fliplr(upLPoint) fliplr(dWindow)];  
        % 绘制边界框
        rectangle('Position', box, 'EdgeColor', [1 1 1], 'linewidth', 2); 
         
        % 如果不是第一帧，计算区域的标签（轨迹关联）
        if i ~= 0  
            overlap    = 0; 
            smax       = 0;  
            imax       = 0; 
            % 查找与上一帧区域的最大交集
            for r=1:length(Box1_save)
                if rectint(box, Box1_save{r}) > smax 
                    overlap    = 1;
                    smax     = rectint(box, Box1_save{r}); 
                    imax = Idx1_save(r);
                end
            end
             
            % 若无交集，则分配新标签
            if overlap == 0 
                if j ~= 1
                    imax = max(max(Idx1_save), max(Idx2_save)) + 1; 
                else
                    imax = max(Idx1_save) + 1;
                end
  
            end
            % 更新当前帧的边界框和索引列表
            Box2_save{end+1} = box;  
            Idx2_save(end+1) = imax; 
        else 
            Box2_save{end+1} = box; 
            Idx2_save(end+1) = j;  
        end
        % 更新last_fr中动态轨迹的质心信息
        info_save{rem(i,Lsmth)+1}(end+1) = I_reg(inds(j)).Centroid(1);  
        info_save{rem(i,Lsmth)+1}(end+1) = I_reg(inds(j)).Centroid(2);
    end
161