Reaearch On The Target Tracking System Based On Mean Shift Algorithm

In recent years,target tracking technology is becoming an increasingly important research directions in the field of computer vision.The improvement of public safety system provides a good hardware platform for the study of security intelligence monitoring system.At traffic junctions, public places,residential communities and other scenes,if you want to monitor the target,the traditional monitoring system requires monitoring staff kept watching the monitor screen while the application of target tracking system enables monitoring system has a higher intelligence. It saves resources and manpower costs and it raises the overall efficiency of the analysis and real-time processing of video data.This paper designs and implements a realistic target tracking system. After manually selecting the target, the system will extract the target feature information, real-time processing algorithm and automatically adjusting the angle of the PTZ. Within the range of the camera,target can be observed and tracked.During the target tracking process,the system uses the improved Meanshift algorithm.It is an effective method to enhance the stability and real-time.This paper also describes the design and the implementation of each module of the system. Using visual studio 2010 software systems to implement the software part of the system.Through indoor and outdoor video capture and retrieval of real-time monitoring to test and verify the designed system.The experimental results show that,under certain conditions,the system can accurately track the target.It has strong adaptability and practicality.The achievement of target tracking system provides a higher level experience for video surveillance research.