Social Group People Relationship Found And People Search Based On Video

Visual surveillance in dynamic scenes, especially for humans and vehicles, is cur-rently one of the most active research topics in computer vision. It has a wide spectrum ofpromising applications, including access control in special areas, human identification at adistance, crowd flux statistics and congestion analysis, detection of anomalous behaviors,and interactive surveillance using multiple cameras.With the large-scale use of surveillance cameras on the social scale, the amount ofdata has become more and more huge. The traditional manual monitoring methods be-come unlikely. So it is becoming an inevitable trend to use data mining method for mininguseful information in a large number of video information. In monitoring scenarios, theuser have to take a long time watching the surveillance video to find the interest event, sothis process is quite time-consuming, and the users have to spend a lot of money to hiresecurity staf. Intelligent video surveillance field is mainly the application of computervision technology to solve the above problems, automaticly test out special events in thelong video monitoring.Based on this, this paper proposes a new method of People Searching. The mainidea is to take into account the people’s group features in real life, we use methods of datamining to dig out the group features, thereby reducing the search space of the target, inorder to improve the accuracy and efciency of the search. And through the experimentsprove the efectiveness of this method.