关键词: Mean shift, Kalman滤波, 目标跟踪, 遮挡

Abstract: In this paper, a target tracking algorithm was proposed by combining the improved Mean-Shift algorithm with the adaptive Kalman filer. For a selected moving object, frame difference and region growing methods were used to segment target, and the dominant color was extracted. In the tracking process, the initial iterative position was obtained by adaptive Kalman filter in every frame, and the tracking result obtained by the improved Mean-Shift was fed back to the adaptive Kalman filter as the measurement for correction. The estimate parameters of adaptive Kalman filter were adjusted by occlusion ratio adaptively. The experimental results demonstrate that the proposed algorithm can detect and track the moving object consecutively in video and has better robustness to occlusion.

Key words: Mean Shift, Kalman filter, Target tracking, Occlusion