An Efficiency-Improved Genetic Algorithm for Enhancing Challenging 3D Prestack Data Using Nonlinear Beamforming

Summary Nonlinear beamforming (NLBF) is an approach for enhancing challenging prestack seismic data corrupted by complex near surface or overburden. It uses a general second-order approximation to locally describe a kinematic wavefront in prestack seismic data. This approximation’s coefficients are unknown and need to be first estimated from input data before the beamforming process. The corresponding optimization problem is highly nonlinear; thus, it requires an efficient and high-quality solver. In this paper, we introduce the NLBF+eGA method, which uses a recently proposed efficiency-improved Genetic Algorithm (eGA) to estimate nonlinear beamforming operators for enhancing 3D prestack data. To further improve the calculation efficiency, we also introduce a concept of ‘spatial consistency’ to the NLBF+eGA method, which uses already estimated neighboring nonlinear beamforming operators as the initial values for a neighboring set of nonlinear beamforming operators. We demonstrate the success of the proposed NLBF+eGA method using a synthetic dataset and a challenging field dataset.