Self-Adaptive Tracking Method Based on Multi-Scale Dynamic Model
基于多尺度动态模型的自适应跟踪算法

This paper simply introduces multi-scale analysis and target-tracking theory, presents a better solution for Px(t), with theoretical proof of the solution's correctness. Based on multi-scale analytical idea, combing the dynamic system analysis of movement model with wavelet transform method, this paper puts forward a multi-scale self-adaptive fusion tracking algorithm,and gives the concrete implementing steps of this algorithm, with simulation experiment. In the simulation experiment, makes a comparison between the algorithm and the traditional single-scale tracking method. And the theoretical proof and the experiment results prove the algorithm to be both feasible and effective. This algorithm adjusts tracking scale automatically according the state of model target. And it also can use detective data effectively and more accurately portray the variation of track. It avoids the disadvantages of single scale and realizes the tracking towards dynamic model. Comparing with the traditional single tracking method, this algorithm more fits the needs of target tracking and has practical value.