Research on application of video anomaly detection in power grid unified video monitoring

With the promotion and application of power grid unified video monitoring platform, the amounts of forward video devices accessing to the platform and the business application systems supported by the platform are increasing. Traditional monitoring technology lagging today cannot meet the requirements of real-time and intelligent monitoring in power production, safeguard and other fields due to defects such as relying on human intervention, unsatisfying monitoring efficiency and overwhelming all-weather workloads on monitoring personnel. This paper expounds the design and development of an intelligent video analysis and management system supported by video anomaly detection technology, based on the existing video resources of the unified video monitoring platform. The system is capable of real-time analysis on abnormal events in videos, also with real-time processing capability and high detection accuracy rate according to the test results, and thus reliable for safe operation of the electric power system.