Research On Video Sequence Target Detection And Tracking Algorithm

With the rapid development of IT technology and the decreasing of all kinds of monitoring equipment costs, the video monitoring system has been widely applied in a lot of fields. Detection and tracking technology of moving targets in video sequences has become a very important research subject.The detection and tracking of video sequence motion target is based on video image analysis, it is a process that combined pattern recognition and image tracking method to detect and track moving object in image sequence, moving target detection is the foundation of the whole target detection tracking system, it is an important part of intelligent video surveillance system that detection to extract moving objects from the video stream and tracking Correctly. However, the actual operation of the scene complexity, interference of environmental noise and the limitation of the algorithm adaption, etc. have brought great difficulties to research work. So, many scholars at home and abroad are committed to moving object detection and tracking research, putting forward many optimization improved algorithms. In this paper, on the basis of predecessors’ research on the video sequence images of moving target detection and tracking, I make some improvements. The main researches of mine are as follows:(1) A new method that combines the continuous frame difference method with the background difference method was presented. The method firstly used continuous frame difference method to obtain continuous difference images. Then a linear adaptive filter or a nonlinear median filter was adopted to obtain the background image and get their subtraction. Finally a threshold segmentation technology was used to enhance the detected moving target. It overcome the disadvantage of the background difference method and the frame difference method and improve the accuracy. Experiment results demonstrate that the proposed algorithm can effectively avoids the phenomenon of the "missed" and "false alarms", etc. and the detection efficiency and accuracy are improved greatly.(2) Proposed an algorithm which combined edge detection algorithm with continuous frame difference method for video moving object detection. The method detects edge on the images of continuous frame difference to fill and expansion the results further. Lastly, I increased the color information of target for detection. Also, it is helpful to track in the field of video surveillance and has great value in practical application.(3) Studied the application of Mean Shift algorithm in video target tracking technology, put forward the improved adaptive Mean Shift moving target tracking algorithm, which combined the adaptive detection of frame difference for sequence images with Mean Shift tracking algorithm, it can adaptively detect and effectively track moving targets. The window width is unchanging in target selection of Mean Shift algorithm is maked up. And it can detect and track the motion target accurately. The algorithm is verified effective through the experiment simulation in the moving target tracking.