3D SRME based on joint regularization and sparse inversion

Surface related multiple is an important problem in marine seismic data processing. Now the 2D wave-theory-based surface related multiple elimination (SRME) is a primary tool to solve this issue. But there are great errors when processing 3D data by 2D SRME because it does not consider the multiple contribution of cross-line direction. The 3D SRME algorithm can be obtained by extending the 2D SRME to 3D space. However, the surface related multiple cannot be predicted by 3D SRME beacause of the streamer feathering and the large cable interval for current 3D marine acquisition. This paper presents a new 3D surface related multiple prediction method that combines regularization with sparse inversion to solve this contradiction between the field data and the 3D SRME algorithm. Using data regularization and de-regularization, this method makes the data distribution satisfy the prerequisites of 3D SRME, and using sparse inversion to eliminate the adverse effect of sparse sampling at cross-line for multiple prediction. The 3D field marine data example proves that this 3D SRME method is effective and feasible.