The Research Of Time-Frequency Analysis Method Based On Hilbert-Huang Transfrom

The major research object of time-frequency analysis method is nonstationary signal, itcan describe the characteristics of frequency spectrum and other statistics variation time.There are some time-frequency methods in recent period, such as short time Fourier transform,wavelet transform, etc. They all play an important part in processing nonstationary signal, butthey still have shortcomings, time resolution and frequency resolution restrict each other, andthe frequency resolution does not achieve the desired results.For the problem of resolution which exists in using time-frequency technology, thispaper studies new time-frequency method, combined with the characteristics of seismic data,develops a high resolution time-frequency method based on Hilbert-Huang transform, andobtains a decomposition method which is suitable for decomposing seismic signal.Hilbert-Huang transform includes two steps: Empirical mode decomposition (EMD) andHilbert spectrum analysis. Firstly, an arbitrary non-stationary signal is decomposed into alimited number of Intrinsic mode functions (IMF) by EMD, then get each IMF component'sHilbert spectrum by Hilbert transform, lastly, get the original signal's Hilbert spectrum bygrouping all of IMFs'Hilbert spectrum.This paper centers on the Hilbert-Huang transform, and then studies the transform itselfand the application of time-frequency technology which based on Hilbert-Huang transform inprocessing seismic data. In the aspect of Hilbert-Huang transform itself, study an EMDdecomposition method which is suitable for decomposing seismic signal. According to thecharacteristics of seismic signal, design the endpoint processing algorithm. At the endpoint,do the spline interpolations according to the change trend of near point of end point. Thismethod suppresses the endpoint effect effectively. In the aspect of application research,extract seismic data attribute using the time-frequency technology which based on HHT.Contrasting the sedimentary cycle analysis which based on logging data with the sedimentarycycle analysis which based on the seismic data nearby wells after HHT, it verifies theeffectiveness of HHT in classifying sedimentary cycle. Applying HHT on the seismic datawhich are original data and data after improving resolution, contrasting the results, there are frequency differences between them. So HHT can be used to classify the resolution of seismicdata. In addition, this paper discusses the possibilities of seismic data de-nosing and highresolution process based on the characteristics of seismic data after HHT. According to thecharacteristics of surface wave after EMD, design the algorithm of eliminating surface waveand achieve good result.