基于自投影和灰度检索的视频帧中异常行为检测

Anomalous Behavior Detection in Video Sequence Based on Self-Casting Histogram and Gray Histogram

针对智能监控系统,提出了一种基于运动目标灰度直方图和自身投影直方图的检索匹配方法,能够快速实现视频序列中行人的运动方向异常检测。该方法结合目标的灰度直方图和自身投影直方图在人群中快速检索匹配目标,采用目标质心运动历史记录表连续记录目标质心和运动方向,通过比较各个目标的运动方向找出运动人群中的异常目标。实验结果表明,引入目标的自身投影直方图,比只利用灰度图的灰度信息有更高的检测准确性,同时历史移动记录表可完全胜任运动目标信息记录的任务。该方法计算量小,同时利用记录质心的移动速度能实时对目标的运动情况进行预测,对运动目标的相互遮蔽有一定的鲁棒性。

For the intelligent video surveillance system,a motion object retrieval match approach is proposed,combining with the gray histogram and the self-casting histogram.It can rapidly detect an object with abnormal direction of motion.The method uses the feature combined with the gray and self-casting histograms to detect and match the object among crowds.And it uses the motion history record list of object centroid to continuously record the centroid of object and its motion direction.Besides,it compares the motion direction to find the abnormal object among moving crowds.The experiment result shows that compared with the method only employing the information of gray histogram,the accuracy of detection is improved after introducing object self-casting histogram,and the motion history record list is fully qualified to record the motion information of moving objects.The method has small amount of calculation and good robustness against objects covered by each other during their movement by recording the speeds of centroid motion.