Study Of Pedestrian Detection And Tracking In Intelligent Video Surveillance System

Intelligent video surveillance is a emerging research in the field of computer vision. This technology automatically analyses sequenced images from video to locate , track and recognize the action of targets in the scene monitored, based on which it makes the related analysis and discrimination in target behaviors, and timely response to the abnormal behavior can be done in the realization of daily management. Pedestrian detection and tracking is the core content of intelligent video surveillance technology, so the study of relevant algorithms is of great significance for the improvement to the system performance.Based on the familiarity and mastery of the relevant digital image processing theory and technology, this paper makes analysis and comparison of the commonly used target detection and tracking methods and works on pedestrian detection and tracking in different conditions.The main research works in this thesis are as follows:(1)Using the moving camera to track people causes size change of target in the dynamic background, which makes it more difficult to track people. To solve the problem,this thesis proposes an improved pedestrian tracking algorithm which synthesizes particle filter and histograms of oriented gradients (HOG) detection and realizes pedestrian track in dynamic background. The algorithm takes the particle filter as the tracking framework, and uses the improved small-scale HOG method to realize pedestrian tracking for the small scale. Also, it identifies the target area according to the result of HOG detection, and constantly modifies particle sampling. Compared with the traditional particle filter algorithm, the simulation results show that the proposed algorithm can track pedestrian target whose size is changing in the dynamic background more accurately and efficiently.(2)The social security video monitoring system is introduced as an important application of Intelligent video surveillance. After analyzing the framework of its imaging processing technology, we complete the design of the pedestrian detection and tracking module in view of the actual monitoring conditions and requirements. The VC++ develop platform and OpenCV library are used for the functional realization of the image extraction and pretreatment, target separation, target classification, feature extraction, pedestrians tracking, and so on. Experimental results demonstrate that this module can deal with pedestrian target in common static background ,which meets the preliminary requirement of the practical application...