对采样数据序列进行时频分解法的改进

经验模态分解法(EMD变换),能够从采样数据序列中提取不同时间尺度波动曲线,以达到时频分解的目的。文章对经验模态分解法的重大改进,主要体现在零平均线的求取上,它无需求上、下包络线,而直接把N点原始数据按时间尺度分段压缩,再用3次样条函数光滑插值到N点,并以此作零平均线,从而得出分层模态分解,改进了Harr小波分解。

The Sampled Data Sequences of Time-Frequency Decomposition Method Improvement

experience mode decomposition method (EMD transform), to be able to extract different time scales from sampled data sequence fluctuation curve, in order to achieve the purpose of time-frequency decomposition. This article is a major improvement for EMD transform, mainly reflected on the zero average line to calculate, it has no demand for upper and lower envelope, It compress the various pieces of data. It interpolates reach to the N points smoothly with a cubic spline, and as it for the zero average line. By this method the hierarchical mode decomposition is acquired. This paper also improves the Harr wavelet decomposition.