经验模态分解中IMF分量判定准则方法实验研究  

Experimental Study on the Method of IMF Component Judgment Criterion in Empirical Mode Decomposition

作　　者：郑健 ZHENG Jian(Railway Engineering Consulting Group Co.,Ltd.,Beijing 100055,China)

机构地区：[1]中铁工程设计咨询集团有限公司,北京100055

出　　处：《测绘地理信息》2021年第3期33-37,共5页Journal of Geomatics

基　　金：北斗卫星技术在精密工程测量控制网中的应用研究项目(研2016-8)。

摘　　要：研究了经验模态分解(empirical mode decomposition, EMD)的原理,将自相关函数法、能量值法、频谱分析法、累积均值法4种本征模态函数(intrinsic mode function, IMF)性质判断准则应用于EMD分析,并以GPS桥梁动态监测数据为例进行了实验分析。实验结果表明,根据不同判断准则得出的结果不同,累积均值法和能量值法由于能够直接定量分析,所以可直接判定。自相关函数法和频谱分析法则需要结合图像分析,不能直接判断。因此,实际应用中要综合判断。The principle of empirical mode decomposition(EMD) is studied, and four intrinsic mode function(IMF) properties judgment criteria are applied to EMD analysis, including autocorrelation function method, energy value me-thod, spectrum analysis method, and cumulative mean me-thod. The GPS bridge dynamic monitoring data are taken to carry out experimental analysis. The results are different when the judgment criterion is different. The cumulative mean method and energy value method can be directly determined because they can be performed quantitative analysis directly. The autocorrelation function method and spectrum analysis method need to be combined with image analysis and cannot be judged directly. So comprehensive judgment is required in actual application.

关 键 词：经验模态分解 本征模态函数分量 判定准则 GPS动态监测数据 结果分析 

分 类 号：P228[天文地球—大地测量学与测量工程]