Crosshole Radar Traveltime Tomography by Occam Inversion Technique
E.E. Baksi and S. Kadioglu
Event name: 10th International Workshop on Advanced Ground Penetrating Radar
Session: Imaging and Tomography
Publication date: 11 September 2019
Info: Extended abstract, PDF ( 666.5Kb )
Price: € 20
The aim of this paper is to present two traveltime tomography algorithms using Occam inversion for crosshole radar method. First, A two dimensional (2D) eikonal equation has been solved by using second order finite difference approximation to compute first arrival traveltimes. Second, a conjugate gradient least-squares (CGLS) inversion, which has been used before for crosshole radar data and Occam’s inversion which have been applied to first arrivals of crosshole radar data to determine slowness distribution and obtain subsurface model. The Jacobian matrix has been constructed by using a finite-difference approximation with the perturbation of slowness for the first iteration, and Broyden's method for later. We tested CGLS and Occam’s inversion methods for the first arrival traveltime tomography of the crosshole radar with second order solution of the eikonal traveltime equation on four synthetic data sets composed from simple to complex subsurface velocity models. The tomography results were stable and convergent. According to the obtained model imaging results, the Occam’s algorithm produced better results than the CGLS inversion.