Seismic inversion is a non-uniqueness problem due to the band-limited wavelet and seismic data. To solve that problem, some prior information is introduced to make a constraint, such as the sparse layer method. It uses the sparseness assumption based on L1 norm to obtain a high resolution result. The Primal-Dual interior method for convex objectives method (PDCO) is usually employed to implement the inversion. However, it brings difficulties in choosing the value of parameter . In this paper, the spectral gradient-projection method for minimization L1 norm (SPGL1) is employed to do the sparse layer inversion. It addresses the PDCO problem by finding an optimizing automatically. Some numerical examples are presented to study the performance of this method. The application result on real seismic demonstrates the potential of this method for deriving high resolution inverted result correctly and efficiently.
Presentation Date: Wednesday, October 19, 2016
Start Time: 1:55:00 PM
Presentation Type: ORAL