Electromagnetic infrastructure monitoring: The exploitation of GPR data and neural networks for multi-layered geometries

In this paper, an inversion ANN-based algorithm for the estimation of geophysical properties (i.e. thickness and permittivity) of subsurface layers in stratified geometries is presented. The basic procedure for the analysis of GPR scans of single subsurface layers placed over a uniform background recently proposed by the authors has been here extended and inserted into a general framework where each stratum is recursively processed.