A Hybrid approach to identify event detection in video Surveillance

Abstract

Event detection in video surveillance has an important task nowadays. Video surveillance will give a continuous monitoring facility to observe normal and abnormal activity in real-time. The current century without the interaction of human surveillance will do but their challenge to identify the events in this paper survey of methods was discussed with their possible successful scenario. The System aims to identify human behavior in particular public premises, by analyzing live video feed from CCTVs. The main focus is that, as a human pedestrian walking on the road may have some continuously changing habits which may affect algorithm working on event detection, and such unforeseen behavior may be flagged anomalous by system wrongfully. In a unified approach for abnormal behavior detection and group behavior analysis in video scenes, current approaches do either use trajectory-based or pixel-based methods. The main objective of to predict the abnormal activity with clear evidence so we can propose the hybrid (integration ) architecture with working conditions.