NOISE ATTENTUATION USING A DIPOLE SPARSE TAU-P INVERSION

摘要

An input seismic data set is obtained (202), and a first inversion is conducted on a first frequency filtered input seismic data set across all slowness values in the plurality of slowness values using an inversion matrix with dipole modulation containing a plurality of Tau-p operators and a plurality of dipole terms to transform the first frequency filtered input seismic data set from a time domain to a Tau-p domain having a plurality of first inversion Tau-p coefficients (208). Each Tau-p coefficient associated with one of the plurality of slowness values. The first inversion Tau-p coefficients are used to identify a subset of slowness values from the plurality of slowness values (210), and iterative sparse Tau-p inversion using the identified subset of slowness values is performed (211). The input seismic data set is modified to generate a noise attenuated seismic data set based on the iterative sparse Tau-p inversion (225).