Time-Varying Seismic Wavelet Estimation Based On The Improved Spectral Modeling In Time-Frequency Domain And Spearman Correlation Coefficient

Seismic wavelets play an important role in seismic data processing and the result of seismic data processing is affected by its accuracy.Due to the energy attenuation and phase velocity dispersion,seismic wavelet is time-varying.The main frequency of seismic wavelet decreases and the phase distortion is time-varying.At present,the accuracy of seismic wavelet estimated needs to be improved.In order to extract the time-varying wavelet with high-precision,methods for extracting amplitude spectrum and phase spectrum were analyzed.This thesis mainly finished the following innovative work.Firstly,in order to effectively solve the problem existing in the time-varying wavelet amplitude spectrum extraction method based on spectral modeling in the time-frequency domain,a time-varying wavelet amplitude spectrum extraction method based on improved spectral modeling in the time-frequency domain was proposed.In order to effectively reduce the energy diffusion of the time-frequency spectrum of seismic records,the synchrosqueezing modified S transform was used to extract the time-frequency spectrum of seismic.By extruding the energy of the time-frequency spectrum,the energy is gathered at the real frequency and the extraction accuracy of the time-frequency spectrum of seismic was improved.In order to solve the problem that polynomial order needs to be determined artificially,an evaluation function describing the quality of wavelet amplitude estimation is established and the determination of the order of polynomials is realized by comparing the value of the evaluation function.The validity of the method was verified by simulation experiments and real seismic data processing.Secondly,in view of the problem that the wavelet amplitude spectrum shape is limited in traditional methods and the accuracy of estimated wavelet spectrum is low,considering that spearman correlation coefficient is suitable for short data,a method for extracting time-varying wavelet amplitude spectrum based on spearman correlation coefficient was proposed.The autoregressive moving average model was used to parameterize the seismic wavelet and spearman correlation coefficient was used to describe the independent characteristics of reflection coefficient series.The parameters of the wavelet model were obtained under the principle of maximum independence of reflection coefficient and then the wavelet amplitude spectrum was obtained.The simulation results showed that this method has higher accuracy than the time-varying wavelet amplitude spectrum extraction method based on spectral modeling in time-frequency domain.Finally,in order to solve the problem that the phase spectrum extraction accuracy of the time-varying wavelet phase extraction method based on phase-only filter is low in the case of low signal-to-noise ratio,a time-varying wavelet phase extraction method based on modified phase-only filter was proposed.The arithmetic average filtering method was used to filter seismic and the weighting function was set according to the results of post-processing.The seismic was replaced by the product of weighting function and seismic in zero-phase discriminant criterion and the influence of noise was weakened.The problem of low accuracy for phase spectrum extraction under low signal-to-noise ratio was solved.The simulation results showed that the modified phase-only filter method has higher accuracy of phase spectrum extraction than the phase-only filter method.