An efficient algorithm for large-scale joint inversion of gravity and magnetic data
عنوان دوره: بیستمین کنفرانس ژئوفیزیک ایران
کد مقاله : 2008-NIGS
نویسندگان
svatan@ut.ac.ir
چکیده
An efficient algorithm for large-scale joint inversion of gravity and magnetic data sets using Gramian coupling constraint is developed. The global objective function is formulated in the space of weighted parameters, but the Gramian constraint is implemented in the original unweighted space. This provides more similarity between reconstructed models. It is assumed that measured data are obtained on a uniform grid. Then, the sensitivity matrices exhibit a block Toeplitz Toeplitz block (BTTB) structure for each depth layer of the model domain, and both forward and transpose operations with the matrices can be implemented efficiently using two dimensional fast Fourier transforms (FFT). The regularized reweighted conjugate gradient (RRCG) algorithm, which relies only on matrix vector multiplication, is then used to minimize the objective function. Application of the RRCG algorithm in conjunction with the BTTB structure of the sensitivity matrices leads to a very fast methodology for large-scale joint inversion of gravity and magnetic data sets. Numerical simulations and real data application demonstrate the efficiency of the presented joint inversion algorithm.
کلیدواژه ها
Title
An efficient algorithm for large-scale joint inversion of gravity and magnetic data
Authors
saeed vatankhah
Abstract
An efficient algorithm for large-scale joint inversion of gravity and magnetic data sets using Gramian coupling constraint is developed. The global objective function is formulated in the space of weighted parameters, but the Gramian constraint is implemented in the original unweighted space. This provides more similarity between reconstructed models. It is assumed that measured data are obtained on a uniform grid. Then, the sensitivity matrices exhibit a block Toeplitz Toeplitz block (BTTB) structure for each depth layer of the model domain, and both forward and transpose operations with the matrices can be implemented efficiently using two dimensional fast Fourier transforms (FFT). The regularized reweighted conjugate gradient (RRCG) algorithm, which relies only on matrix vector multiplication, is then used to minimize the objective function. Application of the RRCG algorithm in conjunction with the BTTB structure of the sensitivity matrices leads to a very fast methodology for large-scale joint inversion of gravity and magnetic data sets. Numerical simulations and real data application demonstrate the efficiency of the presented joint inversion algorithm.
Keywords
Gravity, Magnetic, Joint Inversion, large-scale, Gramian Constraint, BTTB structure