Theoretical Research Of Marine Multi-source High Fidelity Data Acquisition And Deblending

(As a high-efficiency and high-density seismic acquisition technology,the multi-source blending acquisition technology has the advantages of improving the image quality,improving the efficiency of the collection operation and enhancing the illumination of the underground media,but it is subject to factors such as the excitation conditions of the sea source and the exploration cost.The industrial process of marine multi-source blending acquisition is still in the initial stage.In order to realize the transition from theoretical research to actual acquisition,the paper systematically studies the mixed acquisition technology of marine multi-source.This paper mainly explains the basic principle of mixed mining of marine multi-source,the method of separating mining data and the method of reconstruction of missing gathers.In view of the problems such as insufficient illumination and missing gathers in conventional exploration.From the multi-source blending acquisition,sparse transform reconstruction of seismic data and other aspects of the corresponding solutions proposed.The feasibility of the method is verified by a simulation example,which lays a theoretical foundation for the commercial development of the future marine mixed source acquisition technology.In this paper,the basic theory and mathematical model of multi-seismic mixed acquisition are explained by comparing conventional single-source acquisition and multisource blending acquisition.This paper explains the concept of two important parameters of SDR and STR,and introduces the concept of mixed source operator.According to different mixed source operators,the multi-source blending acquisition is divided into: "completely undermined mixed source acquisition" and "incomplete lack of mixed source collection",and its corresponding blended wave field separation were studied.Two kinds of marine multi-source blending acquisition and observation systems are introduced,which are based on multi-seismic source acquisition and observation system.Themulti-ship observation system of marine multi-source,the multi-seismic observation system of marine multi-seismic source,The forward modeling of the two kinds of observation systems is carried out,and the advantages of each acquisition are analyzed.Four different coding design methods are explained according to the different delay time excitation order.The concept of "pseudo-deblending" in multi-source blending acquisition is explained from the angle of response to time domain source data.In this paper,we propose a multi-source blending acquisition data separation method based on weighted multilevel median filtering iteration.The pseudo-separation records of the time domain mixed data are converted to the F-K domain for signal-to-noise analysis,and different selection windows are designed to filter the noise threshold.The F-K filter can not completely suppress the aliasing noise in the mixed acquisition data and improve the separation data fidelity.In this paper,based on the design of F-K selection window,the weight coefficient is introduced into multi-level median filter to improve the weighted multi-level median filter,and an iterative algorithm is designed.The model is calculated with the high conventional deblending technology has higher separation efficiency and higher quality than the separation of signal to noise ratio.In order to make the multi-source blending acquisition and separation records have high fidelity,it is necessary to reconstruct the missing gather data while the wave field is separated due to the existence of "incomplete and bad gathers" in the process of seismic acquisition.In view of this,this paper introduces the Shearlet sparse base for seismic data reconstruction,and proposes a method of reconstructing blending data and seismic data reconstruction based on Shearlet transform in the framework of compression perception.In this method,the time-domain transform of the missing multi-source blending acquisition data is firstly processed,and the corresponding sparse representation is obtained by Shearlet transform after pre-treatment.The fast iterative shrinkage threshold algorithm is used to recover the sparse coefficients in the Shearlet field.Reconstruction of seismic defects and separation of blended data.The optimal sparse approximation property of Shearlet transform is used to preserve the residual signal while suppressing the residualnoise,and the iterative algorithm is designed to optimize the separation result.The simulation results show that the proposed method has better separation efficiency and computation efficiency than other conventional deblending methods.