Research On The Algorithms Of Moving Objects Detection And Tracking In Intelligent Monitoring System

Recent years because of its practical value and wide development prospects, intelligent monitoring system be paid more and more attentions. Moving object tracking become one of the hot topics in computer vision area.This paper researched on moving object detection and tracing with a relatively static background from video sequences, developing a moving object detection and tracing intelligent monitoring system, and realized these methods to simulate some kinds of situations in the system.Firstly, discussing common methods of object detection, adopting the background subtraction method to detect moving object with a relatively still background. Proposing a new method of background updating was proposed named adaptive time domain updating algorithm to establish an adaptive model of the background for the video scene, discussing and analyzing the methods of threshold segmentation, to get the moving object, then carry out the image pre-processing task for further study, such as mathematical morphology analysis, connectivity analysis, etc. prepare paid for the further research of tracking.Secondly, this paper studies the tracking method based on CamShift tracking algorithm and Kalman filter motion estimation. The basic principle and steps of object tracking of CamShift algorithm was introduced, the method of extracting color character and the searching process of object tracking was studied , and the robustness and the feasibility was verified.Research in the method of Kalman filter motion estimation, introduced basic principle and application of object tracking, mainly researching the condition of multi-objective tracking, and discussing the solution of occlusion, analyzing the outcomes through experiments.Finally, experiment system and platform was introduced, commonality,feasibility and validity were verified by demonstration.