摘要
当下针对群体性事件监测的智能视频监控成为热点。对于传统的社会力模型和光流法在检测中的缺点,提出一种改进的基于社会力模型和光流场联合的人群群体事件检测方法 ,利用社会力模型寻找场景中的社会交互力的极大值点,并利用光流法计算以极值点为中心的区域的运动方向信息,用熵来描述区域的混乱程度。实验结果表明,本文的算法可以有效的检测出群体性事件。
The intelligent video surveillance of group incidents has become a hot spot. For the shortcomings of social force model and optical flow method in the detection, this paper put forward an improved method of the social force model and optical flow method. Use the social force model to find the social interaction force maximum points of the scene. Compute the direction of motion information of the region of interest. Use entropy to describe the degree of confusion of the region. The experimental results show that the algorithm can improve the performance of group events detection.
出处
《自动化技术与应用》
2015年第8期78-82,共5页
Techniques of Automation and Applications
基金
基于人流量的视频预警系统研制(编号1301b042014))