People Tracking in the Multi-camera Surveillance System

Abstract

This paper is dedicated to people tracking and identification in the multi-camera surveillance system. In the proposed method, each people-image is extracted among each camera and then is labeled with its color vector. Color vector provides a similar probability for each person appeared in different camera¡¦s surveillance frame. By combining the pedestrian¡¦s trajectory with relations among different surveillance areas, people identification can be achieved substantially. The experimental results show that the proposed method can effectively track and identify the moving people for normal situations in the multi-camera surveillance system.