A New Target Tracking Scheme Based on Improved Mean Shift and Adaptive Kalman Filter
Shangbo Zhou, Peng Hu, Kun Li, Liu Yujiong · International Journal of Advancements in Computing Technology · 2012
In this paper, a target tracking algorithm is proposed by combining the improved Mean Shift algorithm with the adaptive Kalman filter. For a selected moving object, frame difference and region growing methods are used to segment target and extract the dominant color. In the tracking process, the initial iterative position is obtained by adaptive Kalman filter in each frame. The tracking result obtained by improved Mean Shift is fed back to adaptive Kalman filter as the measurement for correction. The estimate parameters of adaptive Kalman filter will be adjusted by occlusion ratio adaptively. Experimental results indicate that the proposed algorithm can detect and track the moving object consecutively and effectively in video and have stronger robustness for occlusion.