Buried-object detection and imaging through innovative processing of GPR data

An innovative two-dimensional (2D) inverse scattering (IS) approach for processing Ground Penetrating Radar (GPR) data is presented to retrieve the electromagnetic characteristics of a buried domain. The developed GPR-IS approach exploits a multi-frequency (MF) strategy to deal with the wideband nature of the available measurements and a multi-resolution (MR) scheme to reduce the ratio between problem unknowns and informative data. Moreover, a customized Particle Swarm Optimizer (PSO) is exploited in order to overcome the limitations of deterministic approaches in finding the global optimum of the arising MF cost function, which is characterized by a high density of local minima (i.e., false solutions).