Video-based The Pedestrian Traffic Statistics Technology Research

Pedestrian flow detection statistical technique is the research focus in the field of computer vision, pedestrians detection statistical gain flow information for each public effective management provides important data sources. This paper obtain video image sequence through monocular camera, video sequence of static background about human movement target detection and tracking method research, in order to realize the detection, positioning and tracking, at pedestrian statistical technique, motion detection pedestrian targets detection and tracking counting algorithm is technology realization and application core.About moving target detection, first introduce some common moving object detection algorithm and their respective advantages and disadvantages. According to background subtraction method, this paper studies how to get background, how to update background model, design based on gray classification extract background method and refresh strategy, and given the shadow detection inhibition and shadow elimination method,finally finished moving targets detection.About moving target tracking, first introduced analyzes some common target tracking algorithm and their respective advantages and disadvantages. According to the paper practical application, put forward the Mean shift target tracking algorithm to tracking target, when moving target motion too fast, or the occurrence of the block, easy to produce the target tracking lost problems, using the Kalman filter combined Mean shift target tracking algorithm to solve these problems, to accomplish moving object tracking.Finally, the paper given sports pedestrian flow statistics flowcharts, and the concrete method of pedestrian flow statistical techniques, the paper do test about different characteristics video sequences, experiment prove this method can accurately detect moving target and carries on the track reach to the purpose of pedestrian flow statistics.