Research And Design Of Target Identification And Behavior Analysis System Based On Multi Channel Monitoring

With the continuous expansion of the campus,the continuous flow of social personnel into the campus,the hidden dangers of the campus problems gradually exposed,the national Ministry of Education and local universities pay more attention to the "safe campus" security project,but the amount can be used for video surveillance system for school is limited,this paper is under the background of this imbalance between supply and demand.This paper is based on the analysis of system functional requirements,then takes a design of multi-channel monitoring system which can be used for target recognition and behavior analysis by the generalized and intelligent way.The intelligent video surveillance system uses two hemisphere network infrared array camera called DS-2CD3310D-1 of Hikvision as the video acquisition,through the Ethernet connection,the PC terminal of the video information for further intelligent active treatment,which includes target recognition,behacior analysis,retrieval and snapshot.In this paper,the intelligent video surveillance system is based on the traditional Vibe algorithm to improve the cavity phenomenon,improve the target detection algorithm,and use the Kalman filter algorithm to achieve the target tracking.On the basis of the target recognition,the intelligent behavior analysis of the moving target is realized by the intrusion detection algorithm and the perimeter alert algorithm.In the unattended state,when the motion target triggers an abnormal alarm,the system can perform autonomous capture for the video frame image containing the abnormal event and store the image locally.In the process of post-mortem investigation,all eligible surveillance video frame images can be retrieved according to the search conditions.After designing and realizing the modular function,the system has been systematically coordinated in the laboratory,and it can be proved by a large number of experiments that the target recognition and behavior analysis of the moving target can be realized with high sensitivity.