Prestack Seismic Gather Optimization Method Research

Pre-stack gathers optimization technique is the preparation of pre-stack inversion,and it’s an important technique in the process of Shale gas exploration and development. Optimized gathers can truly show subsurface features of AVO properties. Improving the quality of seismic data can meet the requirements of pre-stack elastic parameter inversion. For pre-stack seismic gathers rich in random noise, curved noise and having the problem of uneven gathers, This thesis proposes the flow of pre-stack seismic gathers optimization technique and works on the random noise removal algorithm, curved noise removal algorithm, gather flattening algorithm.This thesis first describes the importance of pre-stack gathers optimization technique in today’s new energy exploration, discusses development and research status of pre-stack gathers optimization technique at home and abroad. And then introduces the existing random noise removal algorithm, including Classical noise reduction algorithm, structure oriented filter and so on. For curved noise removal this thesis introduces the traditional noise removal algorithm, Radon domain separation, and then introduces gather flattening algorithm based on the speed and the statistics. It forms a set of suitable methods for the pre-stack gathers optimization studying on the three issues above. The advantages of the methods are verified by theory and practical data applications. The specific work is as follows:1. For pre-stack gathers rich in non-Gaussian noise, this thesis attempts to apply structure-oriented filter based on hybrid optimization to pre-stack gathers random noise removal. When filtering the pre-stack gathers using structure-oriented filter, the mixed norm is introduced to structural orientation filter to remove Gaussian and non-Gaussian noise in the gathers. The experiment results show that the proposed method can preserve structure information better by comparing the results of this method with curvelet transformation noise reduction and structure oriented noise reduction.2. For pre-stack gathers rich in curve noise, this thesis has presented curve noise reduction algorithm based on high precision Radon transform. By adding variable damping factor in the transformation to iterate and update, this method can minimize the Radon transform residuals. In order to speed up the convergence of residual error, damping factor decreased by logistic curve instead of index curve. Setting curvature threshold to separate valid signal and curve noise in the transformation domain, it can improve the quality of pre-stack gathers.3. For pre-stack gathers irregular, The global optimization gather flattening algorithm is proposed and the flow is presented in detail. The fiducial gather is determined by self-similar neighbor Propagation. Setting time window for the seismic gather and selecting the center point as the seed in the window, we can get space-variant correction factor after the global optimization. Then establish the new coordinate system. All data points’ shift can be obtained by interpolation. Finally use the shift to flatten the gathers. Comparisons with commercial software, the algorithm has stable and efficient features.This thesis applies the gather optimization algorithm to different workareas. From the actual effect, it has shown obvious effect in both noise removal and flattening. Compared with the traditional methods, gathers quality is greatly improved.