PRESTACK SEISMIC INVERSION BASED ON ADAPTIVE MIXED-NORM CONSTRAINTS

摘要

Prior information plays a critical role in seismic inversion, which is used to reduce theill-posed problem. Most of the inversion methods assume that the noise obeys Gaussiandistribution. However, due to the diversity of noise in seismic data, it can hardly meet theprior hypothesis. In this paper, a seismic prestack inversion method based on adaptivemixed-norm constraint is proposed to cope with the different noise distribution, andimprove the noise suppressing ability of inversion algorithm in prestack seismic data.First, the noise analysis of actual shale gas is realized through the forward modeling ofwell logging data. Second, the constraints of the L2 norm and the L4 norm are added tothe target function. The new algorithm combines the ability of L2 norm on super-Gaussianand Gaussian noise, and L4 norm on sub-Gaussian noise. It adaptively regulates theweights between L2 norm and L4 norm through Kurtosis. This method improves theadaptive ability of the algorithm to sub-Gaussian, Gaussian, and super-Gaussian noise. Byidentifying the types of noise, the adaptive mix-norm inversion method is used to test themodel, and the prestack simultaneous inversion is carried out in the actual shale gas data.The results show that the proposed method can obtain better inversion results comparedto conventional methods.