Linear tomographic processing of stepped-frequency GPR data

This paper deals with the tomographic inversion of experimental data acquired by a stepped-frequency ground penetrating radar (SFGPR). The experimental SFGPR has been designed for archaeological prospecting and makes it possible to exploit a multiview/multi-static/multi-frequency configuration. The tomographic algorithm is based on a linear model of the electromagnetic scattering to reconstruct the shape of strongly scattering targets. It has been already validated against experimental data for objects in free-space and synthetic data for buried scatterers. Here, we present experimental results for buried objects starting from measurements collected at two test-sites. The first one has been a dedicated outdoor test facility which has allowed to test the experimental setup and the inversion algorithm in controlled conditions; the second one has been an archeological test-site.