基于运动图像序列的异常行为检测*

针对公共重点区域的智能监视问题，研究了一种基于运动历史图像（motion history image,MHI）的行人异常行为检测方法。利用运动图像序列得到的MHI获取视频帧中运动目标的运动方向，由运动方向的变化分类确定人体运动模式和行为是否异常，同时给出相应的实验结果。结果表明，该方法实现简单，具有较好的实时性与鲁棒性,可以作为实时监控系统中异常行为检测的有效方法。

Anomaly detection based on motion image sequence

For intelligent monitoring in the public key areas, this paper researched a novel algorithm to detect pedestrian anomalous behaviors based on MHI.It used motion image sequence for MHI to get the motion directions of human object in eve-ry frames. So it determined whether the movement mode and behavior of the human object was anomalous based on the variation categories of motion directions. And provided some experimental results.The results show that this algorithm is of low computation complexity thus it can be used for anomaly detection in real-time surveillance system.