摘要
Basseville M及 Chou K C等基于二叉树多尺度随机过程提出了一种多尺度数据融合算法。该算法实现简单、速度快 ,易于并行化。本文对此算法做了进一步的研究 ,提出了一种新的多尺度模型构造方法 ,该方法简单 ,便于实现。仿真计算表明 ,采用此方法构造的多尺度状态空间模型对信号有较好的近似 ,完全可以应用 Chou K C等提出的多尺度融合算法 ,滤波效果明显 。
K. C. Chou et al [1] proposed a model for multi scale data fusion of multi scale stochastic processes on dyadic trees. In this paper, we present a method for choosing model parameters. The multi scale process is described by eqs.(4) and (5). We consider the data of every joint in the dyadic tree as the approximation of the signal on a certain scale. The resolution of layer m is twice that of layer m-1 . The multi scale state space model can be described by eqs.(9) through (12). Using the algorithm proposed by K.C.Chou, we can obtain fusion result. Simulation results (section 4) show that our method is effective.
出处
《西北工业大学学报》
EI
CAS
CSCD
北大核心
2000年第2期320-323,共4页
Journal of Northwestern Polytechnical University
基金
国家教委博士点基金资助! ( 97CJ0 80 1)
