Design And Implementation Of Intelligent Video Analysis System For Campus

Nowadays,science and technology affect our lifestyle unconsciously from every aspect.With the rapid development of new technology,our lives become more convenient and comfortable.Intelligent video analysis system,is being used more and more broadly,as one of the most booming industry.The most prominent experience is that our safety can be guaranteed better and better.Currently,campus security problem is being taken seriously by social media and the masses,with the improvement of safety awareness.Intelligent video analysis system can find the abnormity timely and accurately,to promote students' safety.However,most campus video surveillance system is mostly digital and network based traditional video surveillance system.In the aspect of campus security construction,intelligent video surveillance system is seldom used.The system is mainly oriented to this problem.The key technology of intelligent video surveillance system is studied in this paper.Design and implementation of the intelligent analysis system,technology in the process of moving target detection and tracking is improved and the application,based on the moving target detection and extraction of moving objects to track the detected target,improve the work efficiency of security personnel.At the same time,the system applies moving object detection and tracking technology to target intrusion detection.The system can remind the security personnel to deal with the abnormal situation in time.The system in the course of moving object detection,background modeling by modeling method using Gaussian Mixture Model,mainly to the original Gaussian Mixture Model was improved using green image instead of gray image,thereby reducing the amount of computation,improve operational efficiency,while the Gaussian Mixture Model and Frame Different Model the detection results are fused to remove voids and faults,to get the foreground object block complete.In the course of moving object tracking by combining Mean Shift algorithm and kernel correlation filter algorithm to improve the characteristics of target tracking,mean shift method is improved,feature tracking method combined with kernel correlation filter algorithm target position prediction and improved Mean Shift Model HSV color texture,solves the problem of deformation of moving target and scale change,color similarity and occlusion problem.The system mainly through the analysis of the trajectory,for the corresponding movement patterns.Anomaly monitoring,the main target for thespecific area of intrusion detection.The experimental results show that the target detection method.The experimental results show that the target detection method used in the system,can be used to detect foreground objects in the scene,tracking method can be accurately matched to the corresponding target can produce the corresponding object trajectory.Behavioral analysis can detect most of the behavior,and has a good effect on the specific area of intrusion detection.