Research On Abnormal Behavior Real-time Detection In Video Surveillance

In recent years public safety at crowded scenes have been highly concerned.Video surveillance systems are widely used in various places and situations such as street security,traffic analysis and underground transformation safety.Traditional computer vision methods may not meet the demands due to occlusions among pedestrians and illumination poses in a complex crowd dynamics scene.One of the main challenges is to detect abnormal behaviors in densely crowded circumstance both in the time and space domains.The above situation illustrates the challenge and intractability of crowd behavior analysis task.This paper presents an approach for detecting abnormal events in videos by analyzing tracklets extracted by dense optical field.To begin with,this method codes a video as a set of spatio-temporal volumes in order to include the contextual information.After that the dense optical fields is established to extract the dense tracklets.Then the descriptor referred to as Histogram of Oriented Tracklets is utilized to quantize the tracklets in terms of angle and magnitude.Finally,the abnormal behaviors is detected based on probability mechanism using topology of an ensemble of tracklets.Unlike standard approaches that use optical flow,dense trajectories have shown to be efficient for representing videos.The effectiveness of the proposed method was evaluated on public datasets:UCSD,subway and mall.The experimental results demonstrates that this method outperformed former approaches based on the tracklets and optical flow.The abnormal behavior detection in this paper is divided into two main processes.Firstly,this paper presents a novel method to detect salient objects by exploiting the motion information and background information.Then the potential normally motion area is then obtained.Secondly,the algorithm is implemented to determine whether there is a abnormal behavior in the normally motion area and highlight them with light green if the behavior is abnormal.The abnormal behavior detection algorithm is mainly based on the motion information of pedestrians and use the probabilistic mechanism to detect the abnormal behavior.Experiments are implemented on public database like Mall,UCSD and Subway database.Experimental results show that our method is able to detect global and local abnormal behaviors in the presence of shadows,lighting changes interference and the method outperforms others.