Research And Implementation Of The Moving Target Detection And Tracking Technology In Surveillance Video

The video monitor technology gradually turn more mature and people's awareness of safety gets better in recent years,more and more video monitor systems have been applied to daily life.The video monitor equipment is all over the streets of the city,but the traditional human monitoring method can no longer meet the needs of practical applications.Intelligent monitoring technology provides powerful technical support for video monitor to get rid of human monitoring.Intelligent video monitoring technology involves many research fields,and it has broad application prospects in many fields,such as intelligent security,intelligent traffic,intelligent city and so on.The detection of moving targets and the tracking of moving targets are the two main research directions of intelligent video surveillance technology,and the detection and tracking of moving targets is the basic part of the intelligent video monitoring system.It relates to the validity and accuracy of video monitoring information,and it also is a prerequisite for subsequent processing of target behavior analysis and pattern recognition.This paper analyzes and summarizes the development of the intelligent monitoring technology,especially focuses on the detection and tracking of moving targets in monitoring video.The main research contents are as follows:(1)The target detection method of Gaussian Mixture Model is introduced in detail.To overcome the limitations of Gaussian Mixture Model,this paper proposes an improved detection algorithms of moving targets based on three-frame difference and Gaussian Mixture Model.The accuracy,adaptability and real-time of the algorithm are discussed and analyzed through experiments.(2)The paper researches Camshift tracking algorithms in-depth.The common feature matching algorithm is compared and analyzed for the defects of Camshift algorithm,and a improved Camshift algorithm based on ORB feature matching algorithm is proposed.It is proved by experiments that the improved algorithm can meet the basic requirements of real-time performance and effectively improve the tracking effect of moving targets.(3)The experimental platform was built according to the experimental requirements.The module design,process design and system environment of the experimental platform are introduced in detail.The platform interface was designed according to the function of the module.The platform can achieve basic operations such as the detection of moving targets and the tracking of moving targets.