| In this paper, moving object tracking is investigated and special attention is paid to occlusion handling. Object tracking is an important branch of computer vision, which combines advanced technologies and research achievements in image processing, pattern recognition, artificial intelligence, automatic control, compute application and other relative fields. The essence of object tracking is automatic recognizing moving target, deciding the state and the position of the target, and automatic tracking the target. It is difficult because of the unstable image signal and complicated environmental factors. Especially, occlusion problem gradually became the limitation of the practical tracking algorithm. Traditional correlative modal is firstly utilized to solve occlusion after improving its robust performance from several aspects. Then, Hausdorff distance (HD) is deep analyzed to provide more robust occlusion handling. Susan feature pixel detection, advanced matching strategy and new template updating methods are adopted to perfect HD algorithm. A great deal of experiment has illustrated the robust occlusion handling performance. But, this algorithm is only effective when the moving target figure is clear and convenient for feature pixel pick-up. We need develop new algorithm. A kalman filter will be selected to improve correlation tracking performance to solve this problem, which is used to make the template updating adapt to changes in object orientation or illumination. This approach is novel since the Kalman filter has been used in tracking mainly for smoothing the object trajectory. The performance of the Kalman filter is further improved by employing a robust and adaptive filtering algorithm. Experiment has proved the accuracy and reliability of this algorithm.
|
PDF Full Text Request