3D Closed-Loop Surface-Related Multiple Elimination based on GPU Acceleration

CL-SRME has the same theoretical basis as the traditional SRME, but the adaptive subtraction process of SRME is replaced by the inversion concept in CL-SRME, which avoids the problem that the energy of the primary is damaged when the primaries interfere with the multiples in the adaptive subtraction process. With the improvement of computational efficiency, seismic data is expanded from 2D to 3D, and 2D algorithm can no longer meet the suppression requirements of surface-related multiple of 3D data. So based on the principles of 2D CL-SRME and 3D SRME, 3D CL-SRME was developed in this paper, and aiming at high computing cost, GPU acceleration strategy is proposed. The numerical calculation of 3D seismic data shows that the 3D CL-SRME has higher accuracy and wider adaptability, and that GPU parallel computing can greatly improve the computational efficiency 3D CL-SRME, which is conducive to the application of 3D CL-SRME.