基于经验模式分解的山西GNSS时间序列降噪分析  

Noise Reduction Analysis of Shanxi GNSS Time Series Based on Empirical Mode Decomposition

作　　者：平旗 窦立婷 成诚 PING Qi;DOU Liting;CHENG Cheng(Linfen Seismic Monitoring Center Station of Shanxi Provincial Seismological Bureau,Linfen 041000,China;Shanxi Seismological Station of Shanxi Provincial Seismological Bureau,Taiyuan 030000,China)

机构地区：[1]山西省地震局临汾地震监测中心站,山西临汾041000 [2]山西省地震局山西地震台,山西太原030000

出　　处：《测绘与空间地理信息》2024年第12期28-31,共4页Geomatics & Spatial Information Technology

基　　金：山西省地震局预报、科研普通课题(SBK-2211)资助。

摘　　要：基于经验模式分解方法(EMD)分析山西GNSS时间序列。首先,分解原始时间序列得到不同频率固有模态分量(IMF);其次,根据相似性和光滑性规则,选择重构的固有模态分量(IMF),并进行重构;最后,计算降噪前后序列信噪比的变化、降噪后序列占降噪前序列的能量百分比和相关系数3个指标评价降噪效果。计算结果显示,经验模式分解方法结合相似性和光滑性规则能有效分离序列中的噪声信号,得到可靠、高精度毫米级时间序列,为地震预报业务提供更好的服务。This paper analyzes Shanxi GNSS time series based on Empirical Mode Decomposition(EMD).First,we decompose the original time series to obtain the intrinsic mode components(IMF)of different frequencies.Secondly,according to the similarity and smoothness rules,we select the reconstructed intrinsic mode components(IMF)and reconstruct them.Finally,we calculate the change of the sequence′s signal-to-noise ratio before and after noise reduction,the energy percentage of the sequence after noise reduction in the sequence before noise reduction,and the correlation coefficients to evaluate the noise reduction effect.The calculation results show that the EMD method combined with the similarity and smoothness rules can effectively separate the noise signals in the series and obtain reliable and high-accuracy millimeter-level time series,providing better services for earthquake prediction businesses.

关 键 词：经验模式分解方法(EMD) GNSS时间序列 噪声 

分 类 号：P228.4[天文地球—大地测量学与测量工程]