Subsurface Radar Imaging by Optimizing Sensor Locations in Spatio-Spectral Domains

This paper deals with subsurface radar imaging for a two-dimensional scalar setting consisting of a two-layered background medium imaged via a multi-frequency, multi-monostatic configuration. The objective is to reduce data for a subsurface imaging problem without performance degradation by determining the optimal sensor locations in both spatial and frequency domains. In this regard, we present a sampling method that effectively extends the Maximal Projection on Minimum Eigenspace (MPME) algorithm to tackle the semi-discrete inverse problem associated with subsurface imaging. Compared to the state-of-the-art technique, we significantly reduce the required samples for imaging. Numerical and experimental results, the latter concerning a buried water pipe, are reported to demonstrate the effectiveness of the proposed sampling strategy. In particular, for the considered cases, the proposed sampling method shows a data reduction of more than 50% compared to other literature sampling methods.