A Study Of Seismic Virtual Source Signals De-noising

The extensive use of seismic interferometry in actual field exploration has strengthened people’s understanding of the seismic wave propagation and is one of the major advances in geophysics in recent years.The basic idea of seismic interference techniques is to extract useful signals from the original seismic signals and reconstruct the seismic signals by interference.The new seismic signal not only contains the characteristics of the original seismic signal,but reflects some important characteristics that it does not have.Seismic interferometry can be applied to different media of different dimensions.According to the source type,it is divided into active source seismic interferometry and passive source seismic interferometry.This paper mainly discusses the virtual source method in active source seismic interferometry,which converts vertical seismic section(VSP)data into single well seismic section(SWP)data.When people find that traditional ground seismic imaging cannot make a major breakthrough,the detector is placed under a complex overlay.The downhole acquisition system can completely eliminate the establishment process of the near-surface velocity model,which is the weakest link in traditional surface imaging.The virtual source method is an important application in seismic interferometry.The principle is to crosscorrelate the seismic records received by different detectors and convert the VSP data into SWP data.The virtual source method reconstructs the source and detector below the complex upper cover layer,bypassing the complex geological conditions of the near surface,and the reconstructed observation system is closer to the target layer.The virtual source method is a technique for imaging and monitoring geological formations under complex superpositions,and the complexity of the medium is not limited.This paper deduces the theoretical formula of virtual source imaging,and verifies the feasibility and advantages of the virtual source method through simulation data.It can be concluded that it is feasible to use the virtual source method to survey the target area.From the simulated data,the virtual source records obtained by the cross-correlation of the wave field are basically consistent with the seismic records obtained by placing the actual source at the virtual source.The actual post-processing can reflect the actual situation of the underground structure and the model.The information is consistent,so the target area can be surveyed using the virtual source method.The virtual source method has many advantages not found in traditional seismic methods,such as the source being reset closer to the target layer,avoiding the defocus problem caused by incorrect estimation of the covered offset velocity,and thus having better imaging resolution.Based on the virtual source method in the traditional time and space domain,this paper extends to the wavelet domain,that is,the virtual source method in the wavelet domain.Wavelet transform can display the characteristics of signals and noise in different frequency ranges,and can be applied to many fields such as speech,image and geophysics.According to previous studies,2D Gaussian filtering is performed on the part of the virtual source signal containing surface waves using wavelet crosscorrelation.The records obtained after reconstructing the data suppress the surface waves to some extent.The core step of the virtual source method is the cross-correlation of the upper and lower wavefields.Previous literatures have proved that crosscorrelation can be performed in the wavelet domain,which provides a theoretical basis for the denoising method of virtual source signals based on wavelet transform.The degree of correlation between the down-going and up-going waves of different scales determines the output quality of the virtual source.The wavelet threshold denoising method is to transform the noisy signal wavelet transform,and the wavelet coefficient of the effective signal obtained is relatively large,and the wavelet coefficient of the obtained noise is relatively small,and the effective signal can be separated by the threshold.In this paper,the wavelet domain cross-correlation method and wavelet threshold denoising method are used to reduce the random noise in the virtual source signal.The VSP records with random noise are separated up and down,the wavelet cross-correlation is used to obtain the coefficient spectrum of the virtual source signal,and the spectrum is wavelet.The threshold is denoised,and finally the inverse wavelet transform is performed back to the space-time domain.Compared with the virtual source signal obtained by the traditional virtual source method,the reconstructed signal suppresses the random noise to a certain extent and improves the image quality.