Research And Implementation Of Technology Of Target Detecting And Tracking In Moving Background

The moving target detecting and tracking based on video sequences is a major field of computer vision research. It has extensive research and prospect in visual navigation, intelligent control, human-computer interaction, medical diagnostics and security monitoring and many other broad areas. At present, the target detecting and tracking technology in static background has already been widely used. But the carme will move with carme platform in general case,so it’s difficult to get real-time and reliability guarantees,as a result of background motion to the object detecting and tracking very difficult because of the camera movement. So it’s very important to study the target detecting and tracking in moving background.After analyzing the technology of detecting and tracking of static background,this thesis major study in global motion estimation and background motion compensation.First of all, median filter are adopted for image preprocessing in order to eliminate the influent of noise.Then, in order to reduce much time from matching the feature blocks, pre-detection method was used to detect the feature blocks according to the characteristics of background motion and the basic theory of image processing. And then,the adaptive road search algorithm(ARPS) was used to get movement vectors of blocks which are going to be matched based on analyzing and comparing of the several block-matching algorithms.After the motion vector has been work out, the Random Sample Consensus (RANSAC) robust estimation and least squares algorithm (LS) to remove the blocks in the target area and those blocks which are matched falsely. And the remain blocks were used to compute the optimal global motion estimation parameters. Then used bilinear interpolation to finish the background motion compensation. Finally,use the background compensation frame substract the reference frame to get the moving target, and Kalman filter is added to track the eyes for the purpose of increasing the speed of target detectionFirstly, this thesis introduces the knowledge of the target detection. Secondly, this thesis focuses on the background motion compensation technology, and certificates its effectiveness through experiments. Thirdly, based on the background motion compensation, using frame difference method which is a traditional and effective method, and combining morphological operation, moving target can be detected. Finally, the experiments results based on MATLAB platform proved that the algorithm proposed in this thesis can effectively detect and track the moving target in moving background and have a good robustness.