Research On Abnormal Behavior Detection And Tracking Algorithm In Intelligent Video Surveillance System

With the rapid development of China's civil aviation industry,the management of airport in safety became much more serious.Based on the traditional video surveillance system lack of active detection to identify abnormal events,a series of abnormal behavior detection algorithm and target tracking algorithm are discussed to provide ideas and reference for the development and implementation of high performance intelligent video surveillance system.The main study content was summarized follows.For detecting abnormal behavior problems,the calculation of the traditional HOFO features is limited to the statistics of optical flow direction.This article the traditional HOFO feature is improved,The new HOFO features includes the optical flow energy.The improved method enhances the ability of describing the image.This article proposes an anomaly detection algorithm based on weighted optical flow energy HOFO feature.The experimental results show that the detection accuracy is improved compared with the original algorithm.For problems of low detection speed and accuracy,this article applies CNN neural network to detect abnormal behavior,and proposes an anomaly detection algorithm based on convolutional neural network.The algorithm does not need to design a feature extractor,the image can be entered directly,and use the method of local perception and weight sharing accelerating the speed of the algorithm greatly.The experimental results show that the proposed algorithm not only accelerates the speed of the algorithm,but also improves the accuracy.In order to solve the problem of tracking failure caused by occlusion,deformation and illumination changes,an object tracking algorithm based on fusion of apparent features and depth features is proposed in this article.First,extracting the depth characteristics of the target area by the using the trained CNN network,and the HSV space color histogram,then get the fusion feature.then,in the framework of particle filter,obtain the optimal position of the target,get the tracking results,finally update the template.The experimental results show that the tracking algorithm is robust..Finally,the abnormal behavior detection and target tracking system is designed and implemented on the Matlab platform,which proves the validity and practicability of the algorithm.