Font Size: a A A

The Research And Implementation Of Intelligent Video Surveillance System Algorithm

Posted on:2013-08-30Degree:MasterType:Thesis
Country:ChinaCandidate:C Y YuFull Text:PDF
GTID:2248330374479955Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
In recent years, people’s safety requirements become more and more intense withthe development of science technology and information. Intelligent video surveillancetechnology plays an important role in security industry and building safe city, safecampus, harmonious society, it is not only used in government, police, finance,transportation areas, but also in industry, commerce, and family, community buildings.The selection of this paper is based on Key science and technology project ofShandong, the name of it is key technical research for video surveillance system. Ourpurpose is to make intelligent analysis for real-time video sequences and video file datareplacing human brain, then analysis monitoring scene and realize some functions, suchas detection, classification, tracking, and abnormal behavior detection of movingtargets.This paper has mainly content as follows:(1) A new detection and tracking algorithm of objects left based on the twoalgorithms is proposed. Firstly judge whether there is movement between twoneighboring frames using temporal difference algorithm, only after we judge that a fewsuccessive frame images are stationary, detect moving used background subtractionalgorithm. In addition, tracking and simple classification for moving targets is given inpaper by MeanShift tracking methods and the characteristics of targets.(2) Some abnormal targets judgment methods based on video are given in thispaper. After detecting moving target, we analyze whether there is object left inmonitoring areas, abnormal items burst into areas or exists in sensitive areas in theprocess of tracking. A couple of videos are given to prove these algorithms.Experiments obtained good judgment effect and demonstrate that the approach whichare proposed on abnormal behavior detecting is easy, exact and fast, so it has certainfeasibility.(3) This paper describes how to design an intelligent video surveillance systemusing Visual Studio6.0and OpenCV1.0which is an open source image processingfunction library based on C/C++. This system mainly realized six functions which aresudden detection, objects left detection, abnormal-burst detection, abnormal items existin sensitive areas detection, vehicle counting, and car/person classification. It can makereal-time monitor about abnormalities in moving scene, as well transfer and store real-time data to local computer providing ex-post analysis for people. Moreover, it canprovide round-the-clock uninterrupted monitoring.
Keywords/Search Tags:target detection, Visual Studio, OpenCV, background subtractionalgorithm, temporal difference algorithm, abnormal targets judgment
PDF Full Text Request
Related items