Design And Implementation Of Abnormal Behavior Detection In Video Monitoring System On University Campus

In recent years,China's education industry has been developing rapidly,which brings higher requirements for all kinds of school work.The construction of campus security environment is one of the important work contents.Through the construction of campus security environment,we must establish a scientific and effective campus security system.Only in this way can teachers and students have a stable teaching and learning environment on campus,and provide strong technical support for campus human security.In view of the scientific and technological,modern campus security work needs,the traditional way to increase human and material resources has little effect.With the development of information technology,the combination of traditional video surveillance system and information technology,the formation of intelligent video surveillance system makes security technology further improved.This intelligent video surveillance has been widely used in road traffic,residential communities and other places,and achieved remarkable results.Campus video surveillance in Colleges and universities is different from the above places in terms of specific surveillance requirements,but it has a strong reference significance for the abnormal behavior of individuals in public places.Especially for the timely detection report of emotional disorder,fighting and alcoholism in campus,it can further assist security personnel in real-time monitoring of campus.Based on the practices related to abnormal behavior and the specific needs of campus video surveillance,this paper studies the trajectory detection of single moving target and the recognition of abnormal human behavior in campus video surveillance.The main research contents and related achievements of this paper include four aspects:(1)The development status of intelligent video surveillance system is sorted out,the related research algorithms are studied and studied,and their advantages and disadvantages are explored.(2)Aiming at the trajectory detection algorithm proposed in previous literatures,combined with the specific trajectory types that may occur in the campus of colleges and universities,in order to improve the general adaptability of the algorithm,a human hovering trajectory detection algorithm based on angular moving objects is proposed,and the feasibility of the algorithm is verified by relevant experiments.(3)By comparing the template matching method with the state space method which can better describe the essential characteristics of human behavior,is selected to identify abnormal human behavior.Then,the methods of extracting human behavior features based on moments and transformations are analyzed,and a method of human behavior recognition based on non-negative matrix factorization(NMF)and hidden Markov model(HMM)with higher recognition rate is proposed.Relevant experiments are carried out to compare the recognition rate of the proposed algorithm with other related algorithms.(4)Integrating the above algorithms,we add some practical functions that may be needed in campus monitoring,such as intrusion detection,residence detection,items left behind and so on,and integrate all functions into a small system.The related algorithms are verified to prove their effectiveness.