ITERATIVE SPARSE DECONVOLUTION USING　SEISLET-DOMAIN CONSTRAINT

摘要

Deconvolution can help improve the resolution of seismic data. We introduce a deconvolution formulation that can arbitrarily select the resolution level of the seismic data by defining a simple squeezing factor. Considering the ill-posedness of the deconvolution problem, some proper regularizations should be added when iteratively solving the deconvolution-related inverse problem. Traditionally used Fourier-domain constraint can be effective only when the seismic data contains linear events. We propose a seislet-domain constraint to regularize the deconvolution problem to deal with the curved events in seismic data. The seislet transform compressed the seismic data along structural direction, and thus can obtain the optimal sparsity. We apply the proposed method to both synthetic and field data examples and obtain encouraging performance.