Noise Elimination Pre-stack In Seismic Data Processing

During the seismic data procession, noise elimination is a very important link. It is not only a serious impact on the effectiveness of other processions, but also affect the interpretation and application directly. Others have done much work, and gained plenty of useful experience, especially noise elimination after stack. However, with the development of oil exploration, the problems to be solved turn esoteric, and geological conditions from the surface to underground is more and more complex. So seismic data processing have to focus on pre-stack. The kind of noise is variety. There are targeted and adapted to every method. Usually, different methods of noise attenuation have different pertinence and adaptability. Without a method is valid to all kinds of noise. Therefore,the noise elimination should be processed in multi-domain.In the paper, firstly, we have discussed the cause, the distributing, the degree of complex and the type of kinds of noise and especially, have discussed the characteristic of these noise in detail. Secondly, we have analyzed the peculiarity difference between noise and signal in multi-domain and amplify their difference in a domain. Then, we can design an arithmetic operator of noise suppression and realize the separation of signal and noise. And these methods of noise suppression have been demonstrated according to the theory of noise generation and characteristic. At the same time, we have studied the relevant processing module and software and realized them. Thirdly, we have analyzed their scopes and characteristics to kinds of modules.Taking use of many modules to noise elimination in multi-domain, we have concluded as follows: When there is a certain difference between signal and noise's apparent velocity, the module of LINOEL of GRISYS is can gain effective results and less bad influence. The module of SFNPRE in GRISYS is an effective method of suppressing single frequent noise. The module of Zone Abnormal noise Processing(ZAP) of OMEGA can suppress noise effectively. To multiple waves, the parabola method of Radon in GRISYS establishes and suppresses them. The module of FKMUL in CGG can also suppress them using attenuations of multi-wave in F-K domain. The effect of RNAB2D and RNAB3D of GRISYS is obvious to random noise attenuation. Then, we have chosen the representative seismic data which include typical noise and processed them by using pre-stack noise suppression technique in multi-domain and their combination. The effect of processing is obviously. Finally, we have analyzed the whole effect to the technique of pre-stack multi-domain noise elimination. We have built the processing flow of pre-stack multi-domain noise elimination and obtained some understanding.