Research And Implementation Of Abnormal Motion Events Detection In Video Surveillance System

With the continuous development of economic and technology, traditional video surveillance approaches have been unable to meet the people for security needs, Especially for banks, supermarkets, airports, metro, important government departments and so on. It has become a growing concerned problem about how to detect abnormal motion events automatically in the monitoring procedure.Abnormal motion event detection studies a video surveillance system technology for detection and tracking of moving targets. We use background subtraction method and TLD tracking technology to detect and track the targets. The system determine whether there are abnormal motion events by match the tracking trajectory with the samples, and alert when an abnormal event is detected automatically. Trajectory similarity measure method was improved with additional analysis of the direction of trajectory during similarity calculation procedure,which can better adapt to events such as abnormal running whose direction of movement change often.The prototype of abnormal events detection was developed,and it was used in school to detect abnormal motion events,which show that the system has a good practicality to detect the abnormal events during the monitoring process.