101. Weak Orthogonal Matching Pursuit with Geophysical Atom Selection
- Author
-
Alessandro Adamo, N. Bienati, and P. Mazzucchelli
- Subjects
Radon transform ,Attenuation ,ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION ,chemistry.chemical_element ,Radon ,Matching pursuit ,Image (mathematics) ,chemistry ,Geomorphology ,Algorithm ,Selection (genetic algorithm) ,Geology ,Energy (signal processing) ,Interpolation - Abstract
Radon transforms are integral operators widely used in seismic exploration for multiple attenuation, velocity analysis and data interpolation. A reasonable assumption about the Radon transform applied to seismic images is sparseness, where the image energy focuses into clusters in the Radon model space. In this paper we apply the Weak Orthogonal Matching Pursuit framework (WOMP) to exploit sparseness and implement a High Resolution Radon Transform: in particular, we introduce a new automatic selection criterion for the Weak Orthogonal Matching Pursuit algorithm based on the physical model underlying the problem (called GeOMP). Results on real data prove the effectiveness of the method over general purpose selection criteria in terms of both reconstruction quality and decreased computational cost.
- Published
- 2014