Multi-iterative tracking method using meanshift based on kalman filter
Jiawei He, Yingyun Yang · 2014
This paper presents a multi-iterative tracking method using meanshift algorithm based on Kalman filter in order to quicken the relatively slow convergence in the original tracking system, on the condition that Kalman filter based meanshift algorithm has been widely used as a methodology of object tracking. The specific number of iteration in meanshift and Kalman filter, which m and n are used respectively as variables, is decided by automatic computer searching from training sequences under constrained conditions, which will lead to an optimal tracking system. The goal of this algorithm which adapt to a variety of tracking environments is to make promotion to the velocity of convergence, accuracy and robustness. The experimental results prove to be efficient, which means the target been tracking can be located precisely.