Frequency-Hopping GPR Prospecting of Sparse Scatterers Through Bayesian Compressive Sensing

摘要

An innovative ground penetrating radar (GPR) inverse scattering (IS) methodology for imaging sparse buried objects is proposed. The developed methodology integrates a customized Bayesian compressive sensing (BCS) solver within a frequency hopping (FH) strategy in order to (i) regularize the IS problem by enforcing sparsity priors on the unknown scatterers as well as to (ii) exploit the multi-chromatic informative content of GPR data. A preliminary numerical benchmark is shown to preliminarily assess the effectiveness of the proposed FH-BCS technique in providing robust and accurate subsurface images.