The Study Of Seismic Decomposition Based On Synchrosqueezing Wavelet Transform

As useful mathematical tool to analyze seismic signals,time-frequency representation methods have been one of the hot issues in the research of geophysics.The synchrosqueezing wavelet transforms(SST,for short)what has been studied in this thesis is a new time-frequency analysis method.The SST has a sharp and clear time-frequency plane,which can exactly describe the non-stationary seismic signals.It has great scientific and practical value to study the characteristic and applications of seismic prospecting using SST.Three different types of model signals tests show that the SST has a higher frequency resolution than the CWT.The inverse SST tests demonstrate that its reconstruction error is acceptable.Then,the thesis mainly study the application of SST in the seismic decomposition.The tests on the Marmousi model show that compared to CWT,more details about seismic and geological body characteristic can be seen from the constant-frequency slice using SST.The results on the gas-bearing reservoir data show that the low-frequency shadow can be observed from constant-frequency slice.Finally,a preliminary study to attenuate the ground roll based on the SST.The ground roll has a very distinct low-frequency,high-energy features in the SST time-frequency plane,which can be well separated from the effective waves.In this thesis,threshold function in time-frequency domain is used to suppress the ground roll effectively.