动态变形数据的EMD卡尔曼滤波降噪模型

De-nosing Method with Kalman Filter and EMD for Dynamic Deformation Data

文章介绍了kalman滤波与经验模态分解的基本原理,定义了经验模态分解的多尺度分解与合成结构。结合kalman滤波与经验模态分解(EMD)理论提出一种新的动态变形数据降噪模型。文章首先对动态变形信号进行EMD分解,选择合适的高频分解模量进行kalman滤波降噪,然后进行经验模态重构获得降噪后的动态变形信号,达到降噪的效果;最后采用单历元解算的变形数据序列进行模型验证,并与小波降噪模型、直接EMD降噪模型对比,视觉评价与量化结果表明此方法最优。

This paper introduces the basic principles of the kalman filter and empirical mode decomposition,defines multi-scale decomposition and synthesis structure of the Empirical Mode Decomposition(EMD).Combination theory of kalman filtering and EMD,this paper presents a new noise reduction model for dynamic deformation data.Firstly,the new method decomposes the dynamic deformation signal with EMD theory,selects the appropriate high-frequency modulus to decompose noise with kalman filter.Secondly,reconstruct using empirical mode to obtain de-noising dynamic deformation signal,to the effect of noise reduction.Finally,model is validated by deformation data sequence using Single Epoch data,and the validation result is compared with Wavelet De-noising Model,traditional EMD De-noising Model.Visual evaluation and quantitative results show that this new method is best.