Study On The Wave Field Separation Of Seismic Based On The Similarity Of Reflected Wave Signal

The wave field separation of seismic is separate different seismic waves each other from seismic wave field.According to the purpose of wave field separation,it can be divided into many kinds,such as the separation of signal and noise,the separation of different kinds signal and so on.Radon transform and ?-p transform are converging the seismic wave to different "points".Then according to the area of signal and noise in the transform domain,the noise and signal can be separated by removing the noise area.F-K transform and other spectrum filtering methods are also based on the difference of signal and noise area in the spectrum to separating noise and signal by removing the noise area.Although it is easy to distinguish the distribution of signal and noise by transforming the signal into different domains,the area between signal and noise still overlaps.The signal will inevitably be included in the cut-off area of noise,or the noise will be included in the reserved area of signal,so that the signal and noise in the wave field can not be completely separated.At the same time,the phenomenon of mixed wave and pseudofrequency will occur.After amplitude compensation and NMO with P-wave velocity,the P-wave signals of each trace at the same time of common-center-point(CMP)gather of blending seismic data and conventional seismic data and the multi-wave and multi-component wave field are the same reflection point,and their amplitudes and phases are similarity.After amplitude compensation and NMO with P-wave velocity,the same time S-wave signals of each trace in the common-conversion-point(CCP)gather are the same S-wave signals at CCP,and their spectrum also has good similarity.According to the statistical principle of stack,the stack result of random noise in the CMP gather or CCP gather tends to zero after NMO,and the stack result is stacking by the signals corresponding to the NMO velocity in each trace and is independent of noise.So,the amplitude and phase of stack trace is also very similar to the corresponding signal amplitude and phase at the same time of each trace in the gather.Therefore,it can effectively remove the noise in seismic data by using stack trace as model trace to filter the seismic traces in CMP or CCP gather.The wave field separation of blending seismic data is based on on the coherence of the signals in the blending wave field,and the blending noise is distributed randomly in the gather in the non-shot space-time domain.According to this distribution difference between the signal and the blending noise,people using different filter method to separate the signal and noise.According to the kinematic characteristics and dynamic characteristics of seismic wave,the separation of multi-wave and multi-component seismic wave field is carried out by filtering in different domain.The separation of multiple wave is mainly based on the difference of velocity,polarity,amplitude,time and period between multiple wave and primary reflection wave.Above all,this paper studies the seismic data of blending wave field,multi-wave and multi-component seismic wave field,wave field with multiple wave,in the space-time domain CMP gather,CCP gather,source gather,common offset gather(COG)and common receiver gather(CRG),and the signal and noise distributions of these gathers in Radon transform,FK transform,Fourier transform and S transform are also discussed.Finally,a seismic wave field separation method based on the similarity of reflected wave signal is proposed.According to the wave field characteristics of blending wave field,multi-wave and multi-component wave field,and wave field with multiple wave,the signal is similarity between seismic trace and stack trace in CMP gather of blending wave field and conventional wave field after amplitude compensation and NMO with P-wave velocity.And the signal is also similarity between the seismic traces and stack trace in NMO with S-wave velocity's CCP gather of the mult-wave and multi-component seismic wave field.The S-transform spectrum of stack trace can be model to an adaptive filter,and applying multi-level filtering to the S-transform spectrum of each trace in the gather.Then,the model iteration by the results of multi-level filtering,and the new model is applying to the original seismic wave field to separating signal and noise,or other type signal.In the process of trial calculation,theoretical seismic data are used to synthesize theoretical blending data and practical seismic data are used to synthesize practical blending data for the trial calculation of blending wave field separation,the multi-wave and multi-component wave field separation and multiple wave separation are calculated by theoretical seismic data and practical data.In order to verify the effectiveness of the proposed method,the experimental data are displayed at the same amplitude level display for the trial data before and after the wave field separation,and the theoretical data before and after the wave field separation are compared by S-transform spectrum.The test results show that the method of this paper can effectively extract the effective reflection signal in the blending seismic wave field,and effectively extract the S-wave field in the multi-wave field,and effectively remove the multiple wave in the seismic wave field.The method of this paper is an adaptive and full band signal extraction method based on the similarity analysis of S-transform spectrum between stack trace and seismic traces.In the process of signal extraction,the spectrum and space-time distribution of the signal will not be damaged,and the phenomenon of mixed wave and pseudofrequency caused by other filtering methods can be effectively avoided.