中文摘要：

在单目视频场景中,针对提高遮挡条件下目标跟踪和计数的效率问题,提出了一种应用于监控系统中的自动跟踪和计数方法.首先,实现了一个动态背景差分模块,这个模块用于光照变化条件下的背景建模;然后通过获取的前景区域的位置和大小来识别前景目标;最后为了提高遮挡条件下的目标跟踪和计数的效率,提出了一种改进的重叠跟踪器来完成目标跟踪和计数.实验结果表明,该方法比传统方法的平均检测率提高了10％,由此可见,该方法能够可靠地跟踪和计数.

英文摘要：

In a monocular video scene,in order to improve target tracking and counting efficiency under occlusion conditions,this paper presented a scheme to automatically track and count people in a surveillance system.Firstly,it employed a dynamic background subtraction module to model light variation and then treated positions and sizes of foreground regions as decision features to identify foreground objects as characters.Finally,in order to improve the efficiency of target tracking and counting under occlusion conditions,it proposed an improved overlap tracker to complete target tracking and counting.On the experiments of tracking and counting people in three video sequences,the results exhibit that the proposed scheme can improve the averaged detection ratio about 10% compared to the conventional work.Therefore,this method can reliably track and count.