Design And Implementation Of Intelligent Video Surveillance System

Intelligent video surveillance system has a wide range of applications in the future, as its core technology is moving target detection and tracking which attracts a large number of experts and scholars at home and abroad. It has made considerable progress in the field of computer vision. It can realize key information extraction and analysis of useful target in intelligent video source, identify abnormal target in the scene, timely warning to the relevant departments, achieve a level of intelligent monitoring without supervision and is error-free.In this thesis, the object is surveillance video, and intelligent monitoring system involves several core algorithms which have been studied. The main contents include the following aspects:Firstly, in terms of target detection, relatively sophisticated algorithms have been analyzed and compared. On the basis of the difference in the background, frame selection method and gray classification method of Pixel gray classified have been improved to make the target detection more accurate. Through in-depth analysis and comparison of the advantages and disadvantages in the several commonly used shadow-removed method, the method combine the threshold processing in graph segmentation and the improved OSTU threshold that have been proposed to replace the gray value to the target. It can not only meet the gray image shadow removal requirements, but also quickly and accurately separated from the target, shadow, background pixels.Secondly, in target tracking, this thesis uses the combination of OSTU and the Camshift tracking algorithm to overcome the shortcomings and the limitations of the gray-scale image. Target tracking stability has been improved by the traditional Camshift algorithm improvement.Thesis designs and implements an intelligent video surveillance system. The main function of the system include the detection of moving targets in gray-scale image, the target shadow removal, target edge detection, target tracking. The system selected target detection and tracking algorithm verified by testing with high real-time and robustness in the multi-target detection and tracking.Finally, this thesis summarizes the work done, analyses the existing problems of current algorithms and looks into the future to make improvements of the system.