A mean shift tracking algorithm based on the current statistical model

摘要

The video target tracking algorithm based on mean shift has been successfully used in vision tracking field due to its real-time characteristic and robustness. However, when the target moves fast with random trajectory, its performance will decrease heavily and even it will lose efficacy. Aiming at the problem, the mean shift tracking algorithm based on the current statistical model has been proposed. In the method the candidate target's movement characteristic is modeled based on the current statistical model at first and it forecast the candidate target location. Then the optimal target location is updated using mean shift algorithm and it is fed back to the tracking filter as the next measurement. Experimental results show that the algorithm possesses the performance of less iteration number, high tracking precision, and high reliability, and has better tracking effect when tracking rapid moving target.