The Research On Group Events Warning Method Based On Video Surveillance

With the development of society and economy, the demand of public security becomes stronger. A large number of video surveillance equipment is being applied at the train station, square, and the government gate areas. Video surveillance technology current stay in the video acquisition and we can’t carry on the intelligent analysis, monitoring task group events still rely on human works. With the demand of the video surveillance system becomes higher, the requirements of intelligent video surveillance system is booming as the same time. How to accomplish the analysis of video sequence of the crowd which obtained from surveillance equipment existed, and early warning of abnormal conditions, are the research focus in recent years.In this paper, the project of "Research on some problems of group events detection method based on video monitoring" is studied, the main contents include:First, group anomaly detection methods were discussed. We used social force model to analyze scene. Estimate the distribution of social force of the crowd through the social force model analysis of abnormal situations, the experimental verification of the validity of the model.Second, crowd abnormal issue based on the scene were studied. A group events detection method combined social force model with optical flow was proposed. Experiments demonstrated that the method improves the accuracy and timeliness of the detection of group events.Third, in the analysis of the moving foreground region extraction method and adaptive scale window area based on Gaussian kernel method. The target area extraction method is proposed to combine the two approaches, and used for group events detection based on video. Experiments demonstrate the effectiveness of this method.