Comparative study between different frequency strategies for relative dielectric permittivity and electrical conductivity reconstruction. Application to near subsurface imaging

This work proposes a comparative study between different frequency strategies for the simultaneous reconstruction of the relative dielectric permittivity and electrical conductivity of the near subsurface. Data on the electric field are generated in the framework of a ground penetrating radar (GPR) configuration. Assessing these two parameters from the data requires solving a nonlinear and ill-posed inverse problem, which is solved iteratively by a regularized Gauss-Newton (RGN) algorithm. Numerical results allow to determine an optimal strategy, for which convergence rate and computation time are reasonable, spatial resolution is improved and the two parameters are very well reconstructed.