基于滤波辅助的PEEMD与Hilbert谱分析方法的高程时间序列特征提取

Feature Extraction of Height Time Series Based on Filter-Assisted PEEMD and Hilbert Spectrum Analysis

利用滤波辅助的部分集成经验模态分解(partly ensemble empirical mode decomposition,PEEMD)与Hilbert谱分析结合的方法建立了一种高程时间序列特征提取方法。首先,采用滤波辅助的PEEMD方法分解高程时间序列中的特征分量,然后应用Hilbert谱分析在频率域上对特征分量进行分析。将该方法应用于中国区域的国际GNSS服务(International GNSS Service,IGS)测站十几年的高程时间序列中,并将滤波辅助的PEEMD的结果与传统的小波分解的结果进行对比。结果表明,该方法准确有效地提取出了IGS高程时间序列中4 a、2 a、1 a周期的特征分量,且滤波辅助的PEEMD分解的结果与小波分解结果一致,可以获得一个与原始时间序列符合较好的趋势项。

We establish a method for feature extraction of height time series by combining filter-assisted partly ensemble empirical mode decomposition(PEEMD)with Hilbert spectrum analysis.Firstly,the feature components of height time series are decomposed by filter-assisted PEEMD method,and then the feature components are analyzed by Hilbert spectrum analysis in frequency domain.This method is applied to the height time series of International GNSS Service(IGS)stations in China for more than ten years,and the results of filter-assisted PEEMD are compared with those of traditional wavelet decomposition.The results show that this method can be used to extract the feature components of 4 a,2 a and 1 a periods in the IGS height time series accurately and effectively,and the results of the filter-assisted PEEMD is consistent with those of wavelet decomposition,and a trend term can be obtained which is in good accordance with the original time series.