Moving Target Tracking Under Complex Background And Technology Research

Moving target tracking can again the target location from the video on time and track object automatically. It is a important technology that has been widely used in the video surveillance, human-computer interface and image compression. In the presence of noise, intensity change, object occlusion and complex background conditions, target tracking is demanded to accurately and effectively get object. So this paper focuses on the detecting the moving target, motion estimation and background compensation. In this paper, several algorithms have been improved to deal with the drawback of Mean Shift tracking algorithm.On the research of image pre-processing, we introduce the methods of reduce noise. For motion Estimation and background compensation, we introduce the block matching and local projection algorithm. Background compensation can change moving background into still background, reduce the difficulty of detecting and tracking moving object. We analyze the dilation, erosion, opening and closing operations in post-processing of image, which have been used to eliminate holes in the image.On the research of moving object detection, this paper introduces three kinds of moving object detection algorithm, combines the background compensation and HOS algorithm to detect target and get more precise resultsAbout moving object tracking, this paper introduces the theory about Mean Shift and verifies the convergence of the algorithm. It points out the advantages and the disadvantages of the traditional target tracking algorithm based on Mean Shift and makes the following improvements for its shortcomings:First of all, the improved algorithm combines the Mean Shift with orientation histogram. Orientation histogram can include the orientation information of pixels. The improved algorithm can increase the tracking reliability and availability, deal with the drawback of getting bad tracking result under the image having little grayscale and texture information.Secondly, this paper combines the Mean Shift with kalman filter that has been used to forecast possible position of target. Mean Shift algorithm searches the final position within the possible area. The improved algorithm effectively solved the problem of occlusion in target tracking.The improved algorithm effectively overcomes the shortcomings of the original algorithm, extends the range the algorithm and improves the algorithm’s accuracy and robustness.