-
oa Four methods of data regularization
- Publisher: European Association of Geoscientists & Engineers
- Source: Conference Proceedings, 71st EAGE Conference and Exhibition - Workshops and Fieldtrips, Jun 2009, cp-129-00035
- ISBN: 978-94-6282-103-3
Abstract
When seismic data regularization is formulated as an inverse problem, it requires mathematical regularization, a method for imposing constraints on the reconstructed data. Mathematical regularization can take four different forms: a differential operator (such as a prediction-error filter or a plane-wave destructor), an integral operator (such as a recursive inverse of prediction-error filtering or a plane-wave constructor), a sparseness constraint in a special domain (such as Fourier or seislet), or a shaping operator. Similar results can be achieved with different methods but at a different computational cost. Using both onedimensional toy examples and seismic field data applications, we compare and illustrate properties of the four methods of data regularization.