一种多尺度数据融合模型的工程实践与相关理论问题

为解决不可重复测量物理量的真值估计问题,提出了一种"多尺度数据融合模型",该模型采用多传感器对同一物理量同时进行测量,对各个传感器的测量结果进行小波变换,并在不同小波尺度域对各个传感器的测量值进行多尺度加权平均,通过逆小波变换得到待测物理量的真值估计。笔者对该模型进行了分析和总结,包括对相关领域研究的评述,在此基础上,给出了数据融合定理及一些重要的推导及结论。

Engineering Practice of a Multi-Scale Data Fusion Model and Its Related Theories

In order to solve the true value estimation problems of unrepeatable measurement physical quantities,a kind of "multi-scale data fusion model" is put forward by the author.This model adopts multiple sensors to measure simultaneously the same physical quantities,wavelet transform of measure results of each sensor has been performed,then in different wavelet scale domain measurement value of each sensor has been weighted and averaged on multi-scale,finally,the real value estimation of the physical quantity to be measured can be obtained by reverse wavelet transform.This paper is an analysis and summary of the multi-scale data fusion model proposed by the author,including reviews of the related research fields,and on the basis of introducing this model,the theorem concerning data fusion and the important derivation and conclusions are given in this paper.