Study On Abnormal Behavior Recognition Based On Video

Video surveillance system as an essential technical means of security,it plays a very important role in all areas such as security protection,and it has been applied to all aspects of life.Filtering out the Invalid information from video surveillance system can not only improve the real-time performance of human behavior,but also can identify the behavior accurately.It effectively reduce the cost of manpower and material resources,while reducing the amount of calculation.This realizes the function of real-time active early warning,the purpose is to solve the traditional video surveillance system afterwards.In the identification of abnormal behavior of the movement of the human body,this paper mainly carried out the following research work.(1)At the time of image preprocessing.First of all,the paper analyzes the overall characteristics of the image and presents the idea of image preprocessing.The paper uses image enhancement and gray level transformation and other methods to remove most of the interference impurities,and it has achieved good results.(2)In the target detection.according to the actual situation of this article,the traditional detection method is analyzed experimentally.In order to solve the traditional single method of interference and incomplete contour defects,this paper proposes a moving object detection based on improved GMM background modeling.The method can effectively remove the external interference and obtain relatively intact contour,it improves the real-time and accuracy of the system,and to a certain extent,the shadow has a better inhibitory effect.(3)When the target is tracked.Different from previous single frame extraction,we choose the representative frame image in the interval fixed frame in this paper.It improves the accuracy of the identification judgment and improves the timeliness of the system.A hybrid moving target tracking algorithm based on Kalman and Mean Shift is used in this paper,this method makes the overall performance of the target tracking more stable and efficient.(4)In the case of abnormal behavior recognition.This paper is different from the traditional single person behavior recognition.The state space method is used to establish the standard behavior database.At the same time,it can effectively improve the timeliness of the system by comparing the behavior of multiple people to reduce the time to match the behavior.This paper also uses the hidden Markov model to identify the human behavior.In this paper,the simulation experiment of the abnormal behavior of the human body is verified,which proves that the algorithm has strong timeliness and accuracy.