Research And Implement Of Pedestrian Flow Detection In Intelligent Video Monitoring System

Intelligent Video Monitoring is attracting more and more attention as an effective measure of security. Meanwhile, pedestrian flow detection has been widely applied to many public places such as bus stations, avenues, and the entrances of the supermarkets. On the basis of the target detection, classification and tracking, this paper introduces the implement of the pedestrian flow detection within a network intelligent video monitoring system.In the matter of. moving target detection, we use the Guassian background modeling method to build our background of the areas under supervision. And a combination of background difference algorithm and time difference algorithm is designed to detect moving target.In the classification of the moving targets, a simplified way to get the interested targets is proposed. The length-width ratio as well as the gradient of the vertical projection histogram are chosen as features to separate the moving people from cars, which forms the basis of the subsequent processing of passengers tracing and flow statistics.In the aspect of pedestrians tracking, first we propose an algorithm called mantle ratio which can automatically select the interested region. Secondly we choose the weighted color histogram of the tracking region as feature. Finally we can track pedestrians under the Framework of particle filtering theory automatically and real-timely.At last, on the basis of the above research, we implement the function of pedestrian flow detection according to the requirements of the intelligent security surveillance applications, which can detect, track and count the moving targets. It also has been performed simple and effective under the real complicated environment.