An innovative Frequency Hopping multi-zoom inversion strategy for GPR subsurface imaging

This work deals with subsurface imaging of buried objects when Ground Penetrating Radar (GPR) is used for gathering time-domain data at the air-soil interface. An innovative inversion procedure based on the integration of a deterministic Conjugate-Gradient (CG) reconstruction procedure and the Iterative Multi-Scaling Approach (IMSA) able to exploit the frequency diversity of GPR data is presented. Here, suitable frequency components of the spectrum of the GPR field data are exploited through a Frequency Hopping (FH) strategy within the proposed IMSA-CG multi-zoom iterative inversion scheme. Numerical simulations show satisfactory results when dealing with scatterers characterized by various dimensions, shapes, and dielectric characteristics. Furthermore, the proposed method demonstrates to be robust against different noise levels and able to provide acceptable reconstruction accuracy although a limited amount of collected data.