Specific Abnormal Behavior Recognition Technology And Implementation In Intelligent Monitoring System

With the rapid development of information technology and people’s attention to social security,video surveillance systems have been widely applied to various fields and industries. But thetraditional video surveillance system still has some limitations, mainly in the traditional monitoringsystem relies heavily on manual analysis and decision-making, traditional monitoring system can’tget real-time information, and in many cases it can only be used after the evidence, can not achievereal-time alarm. Therefore, it is very important to improve the intelligent of the monitoring system.In order to solve this problem, this thesis uses the intelligent ways to recognize the abnormalbehavior for the prison environment, so that abnormal behavior occurring in prisons can beautomatically identified through intelligent video surveillance system, and provide alarms.This thesis investigates the preliminary analysis of video processing and valid judgment foraggregation behavior for multiplayer. It describes some basic techniques of video surveillance, alsointroduces some preprocessing on the Gaussian mixture background modeling and the technologyof foreground extraction. In addition, through analysising target behavior in the dynamic image,abnormal behavior about aggregation can be identified by using the algorithm based on distancevector. After this, this thesis solves the problem of portrait overlap with the way of histograms, thenit present an effective way to solve the aggregate behavior for a special environment—prison. Thispaper is based on the Windows environment, using VC++6.0and computer vision library—OpenCV to implement the specific algorithm and verify the effectiveness of the algorithm.