Design And Implementation About Abnormal Behavior Analysis On Multi Background

With the development of security system,the demand of intelligent and automation is become more and more strong.The ability of traditional security system is only provide the record of the scene have occurred,it have not the ability to know the scene have occurring.So our thesis aim to detect seven abnormal behavior,as regional invasion,ATM stickers and linger etc.besides,our thesis design a system for abnormal behavior analysis.As the requirements of security system,the abnormal behavior analysis that our thesis study have object detection,object tracking and abnormal behavior analysis.In the object detection,the Mixed Gauss model method and Background subtraction are used in our thesis for coping with complex backgrounds,detecting moving target and state target and dealing with light mutations.In the object tracking,the Kalman algorithm and Mean shift algorithm in our thesis are used to track target.We boost the function of iteration search radius adaptive and we adopt many-ways for target matching to reduce the iteration times of Mean shift algorithm,and improve the stability of multi-target.In the abnormal behavior analysis,according to behavior characteristics,we can analysis seven abnormal behavior.In the last,our thesis designed a system for behavior analyzing.At last,after object detection,the precision of detection is prompt by our thesis.By object tracking that our thesis studied,the precision of object tracking become more and more accurately.Our thesis build a system for security.This project is based on security requirements,looking forward to contributing to the intelligent development of security system.