Research And Application On The Method Of Denoising And Regularization In The Mountain Area

In recent years,Complex mountain area have become the main areas of onshore oil and gas exploration in China.A number of high-yield oil and gas wells have confirmed the high exploration potential in this areas.However,the oil and gas seismic exploration in these areas is facing the difficulties of unclear target imaging and inaccurate imaging.As the key link of seismic exploration processing technology,the upgrading and breakthrough of denoising and regularization technology is of great significance to the overall improvement of seismic data quality.This paper studies the new technology of seismic data de-noising and regularization processing in the mountain area.The main research work includes the following contents:1.In view of the problem that the noise is very developed in the mountain area,a regular interference suppression technology based on Cadzow filtering method is developed.Firstly,according to the apparent velocity of regular interference,the seismic data are linearly NMO corrected in the space-time domain,and the high-dimensional Hankel matrix is constructed,the singular value decomposition is performed and the signal is reconstructed by rank reduction.Finally,the regular interference is subtracted from the seismic data to achieve high fidelity de-noising.This method can solves the problem of coupling separation between regular interference and effective signal in prestack seismic data.2.Using the advantage of the fractional Fourier transform can rotate at any angle,a hybrid Cadzow filtering method based on the fractional Fourier transform is formed.Firstly,the seismic data are transformed to the time-frequency domain by using the fractional Fourier transform,and three types of high-dimensional hybrid Cadzow matrices are constructed.Then,the LSVD fast decomposition algorithm is implemented,and the signal is reconstructed by rank reduction.Finally,the effective signal in the seismic record is predicted by using the inverse fractional Fourier transform,so as to suppress the strong random noise and improve the signal-to-noise ratio of the seismic data.3.On the basis of the above-mentioned high-dimensional matrix singular value decomposition method,the mutual information formula is introduced to calculate the mutual information difference spectrum of each component signal and the original signal,and to find the maximum mutation point,that is,the contribution rate of the adjacent two component signals to the seismic signal mutates,which is the boundary point between the effective signal and the noise signal,so as to realize the adaptive selection of the number of effective singular values,The difficulty of selecting effective singular value under strong background noise is solved,and the fidelity of seismic data denoising in foothill belt is improved4.Based on the principle of compressed sensing,the trial function is constructed by using the basis of signal in sparse transform domain,and the evaluation factor of random undersampling is obtained by trial function,which can be used to quantitatively analyze the advantages and disadvantages of different random undersampling methods.Based on this research method,in the design of random undersampling observation system,a large number of random undersampling observation systems can be generated with a certain distribution or multiple distributions,and the optimal random undersampling observation system can be quantitatively evaluated by the optimal evaluation factor,which provides theoretical support for the subsequent data regularization reconstruction.5.Taking the shot point coordinates and detector point coordinates as the dimensions,the generalized distance between any seismic traces is defined,and the data volume to be reconstructed can be indexed and constructed through the generalized distance.Then Fourier transform or curvelet transform is selected as the sparse representation method of seismic data,and the shrinkage threshold iteration method or convex projection iteration method are used to reconstruct the regularized data volume,the seismic data are obtained with high fidelity reconstruction.This method can greatly improve the signal-to-noise ratio of prestack gathers and effectively suppress migration noise.