Full text loading...
-
Curvelet-Based 2-D Inversion for MT Data
- Publisher: European Association of Geoscientists & Engineers
- Source: Conference Proceedings, 81st EAGE Conference and Exhibition 2019, Jun 2019, Volume 2019, p.1 - 5
Abstract
In this paper, we propose a new method for two-dimensional (2-D) magnetotelluric (MT) inversion based on the curvelet transform. Unlike the conventional inversion methods that apply constraints on the model in the space-domain, the method presented in this paper is based on the sparse constraint by the curvelet transform, and we directly invert the curvelet coefficients instead of the model conductivities in the space-domain. The curvelet transform is a multiscale sparse scheme that transforms the model parameters into the curvelet coefficients at multiple scales. The basis function of the curvelet transform is the “wedge base” that satisfies the anisotropic scale relationship (width∝length2) and has the characteristic of arbitrary directivity. Thus, it has the capability to “optimally” represent the edge of the target objects. To achieve a sparse constraint, we use L1-norm of the curvelet coefficients for the inversion. This can help extract the features of target objects more sparsely and get high-resolution inversion results. We compare the results of our curvelet-based inversion with those based on the traditional L1-norm and L2-norm inversions. The experiments with theoretical data demonstrate that the sparse constraint inversion based on the curvelet transform can better reveal the boundaries of target objects.