Multi-frequency GPR data fusion

Abstract

Several data fusion approaches have been developed to maximize both resolution and characterization depth for GPR. In this study, we propose a new physically-based method to merge multi-frequency radar data coming from different antennas. The strategy relies on the removal of antenna effects and the subsequent fusion of the resulting Green's functions in the frequency domain. We validated the approach in far-field conditions using two different antennas operating in the ranges 0.8-4.4 GHz and 2.6-6.2 GHz, respectively. We performed radar measurements above a sandbox wherein a number of objects were buried. The physically-based data fusion provided enhanced radargrams due to the broadening of the spectral bandwidth. Future research will focus on near-field conditions, for which the removal of antenna effects should be performed numerically.