Design And Implementation Of Ptz Camera Based Real-Time Detection And Tracking System For Abnormal Events

In recent years, the technologies of abnormal events real-time detection and tracking in intelligent video monitoring system become a hot research topic in the field of computer vision. The traditional video surveillance system consumes a lot of human resources and cannot complete the analysis of the abnormal events, so the abnormal events real-time detection and tracking technology has substantial research significance and application value.This thesis mainly studies the abnormal event detection and target tracking algorithms and also carries out the analysis and comparison of the algorithms. First, this thesis presents a method of pedestrian detection based on HOG features and support vector machine (SVM) algorithm. Vehicle detection uses Harr feature and Adaboost algorithm. Secondly, a Camshift tracking algorithm based on CENTRIST visual descriptor is proposed, which improves the accuracy of target tracking. We design PTZ camera based real-time detection and tracking system for abnormal events. It uses the PTZ camera to realize monitoring in multi-angle. Through the control of the PTZ camera, this system can prolong the time of target tracking and get more abnormal targets information. In addition, we apply super-resolution image reconstruction technology based on POCS to the system. Experiments on actual monitoring scenario show that this system can obtain ideal effect.