Research On Improved Synchroextracting Transform And Its Application In Seismic Sigal Processing

As an important branch of non-stationary signal processing,time-frequency analysis(TFA)has been one of the hot topics in modern signal processing.The TFA can simultaneous describe the energy density and intensity of the signal at different time and frequency by transforming the one-dimensional signal to two-dimensional time-frequency(T-F)plane,besides,it provides more local feature information so that people can understand the changing characteristics of the spectrum and others.Therefore,it has been widely used in many fields such as bearing detection,voice signal recognition,and seismic exploration.Time-frequency resolution is one of the most important indicators for evaluating the TFA.With the increasing demand for the precision of analysis,the research of high-resolution TFA method has been paid more and more attention,which promotes the improvement of classical TFA method and the proposal of high-resolution TFA method.Synchroextracting transform(SET)is a novel TFA method,which is based on classical TFA methods and can achieve ideal T-F resolution.This method creates a synchronous extraction operator based on STFT to extract the T-F coefficients of the T-F ridgeline on original time-frequency spectra and get a new high-resolution T-F spectrum.SET not only avoids the Heisenberg-Gabor uncertainty principle,but also clearly increase the accuracy of T-F resolution.Furthermore,the SET method has a unique advantage in seismic processing according to its translation characteristics,which indicates that the application of SET method.In this paper,by studying the method of SET,we have improved the method in different perspectives and proposed two methods,one is combining the Synchroextracting Transform and Empirical Wavelet Transform(EWT-SET),and the other is synchroextracting generalized S-transform(SEGST).Besides,we explore the application of those methods in seismic signal analysis.More details are as follows:(1)We have deeply researched the theory of synchroextracting transform,made conclusion and simulations for the source of the idea,the principle and the characteristics of the T-F analysis.Those works pave the way for improving the synchroextracting algorithm and its application in seismic signals analysis.(2)In this paper,we apply SET to seismic signal analysis.Besides,according to its disadvantages of processing multi-component nonstationary complex signals,we combine EWT and SET,named EWT-SET,for time-frequency analysis.The result of synthetic signal verifies that the T-F characterization capability of EWT-SET is better than SET or EWT.The method has been successfully applied to the time-frequency analysis of seismic signals and reservoirs detection.The synthetic model and field data confirm that EWT-SET method can effectively extract the attenuation characteristics affected by the reservoir and is effective in reservoir detection.(3)Based on the fact that the Generalized S-transform(GST)is more flexible and widely used than the Short-Time Fourier Transform(STFT),this paper proposes a new method named synchroextracting generalized S-transform(SEGST).At the same time,the mathematical expressions of the method are detailedly deduced.The synthetic signal shows that SEGST method not only have better T-F resolution,but also be flexible which makes it can deal with both high and low frequency.Furthermore,the SEGST is successfully applied in analyzing time-frequency of seismic signals and thin reservoir identification.In addition,compared and analyzed the results between SEGST,STFT,GST and SET.The results show that the SEGST has a higher identification accuracy.