Research On The Method Of Seismic Wavelet Extraction Based On Higher-order Spectrum Analysis

Using the reasonable and effective way to extract the seismic wavelet accuratelyfrom the seismic data is an important task in the area of seismic exploration. It is thebasic work for the seismic deconvolution, migration, feature extraction andgeophysical interpretation. The application of seismic wavelet estimation using themethod of signal high-order spectrum not only pushed the analyzed method from thetraditional second-order into high order, but also made up for the disadvantage bytheloss of phase information of the second-order statistics method, and the real waveletcan be recovered from the high-order statistics, the resolution of seismic data has beenimproved by this method; the analyzed method of high-order spectrum can effectivelyinhibition the colored noise from the real seismic data, and it has made the effectivewave more prominent. It is incomparable to the conventional method of thesecond-order statistics.This article is based on high-level statistical analysis method and the exactlycalculus process in the seismic wavelet extraction has been thoroughly studied andanalyzed. The high-order spectrum of the signal and its research back ground in thearea of seismic exploration and the research thought which is based on third orderspectrum has been proposed in this paper. Then the basic theory and characteristics ofhigher order statistics have been detailed studied and analyzed, combined with theimportant role of time delay estimation in the seismic exploration, three methods oftime delay estimation has been verified through examples, including time delayestimation methods based on high order spectrum. According to the formingprincipleand theoretical model of seismic wavelet, several common methods about theseismic wavelet extraction have been discussed. According to the drawback of theexisting wavelet extraction method,a method of wavelet extraction through thebispectrum has been introduced and the method has been detailed analyzed. At last thefeasibility about the high order spectrum especially the bispectrum and trispectrumused in the seismic wavelet extraction has been verified by the simulation. It can beseen the seismic wavelet extraction using the method of high-order spectrum don’tneed to make any assumptions about the wavelet phase and the signal noise. And it isan objective method of wavelet estimation. This paper brings the modern signal of high order spectrum analysis applied tothe extraction of the seismic wavelet, reasonable using of high-order spectrumcontains both amplitude and phase information, recovering the wavelet amplitude andphase spectra directly from the seismic recordsand the wavelet is restructured.theachievements obtained in this paper is that the different algorithms are used to recoverphase from the wavelet of different phases., including BLW and BMU,The improvedBMU algorithm solves the problem of phase-jumping by using the method of makingthe tail adjacent phase values linearly related,it has also solved the problem of phasewrapping,and makes the recovery of seismic wavelet phase more accuracy.Using thetraditional comparative analysis of the autocorrelation method and high-order spectralmethod, which fully demonstrates the superiority of the higher order spectral method,it can accurately extract the wavelet. The theoretical modelresults show thathigh-order spectrum to estimate the waveletof high stability, strong anti-noiseability,In view of this research, we can see the potential ofhigh-order spectrumtechnology using in seismic data processing.