GPR Measurements Inversion in Case of Spatially Varying Antenna Polarization

Abstract

In this paper, Ground Penetrating Radar (GPR) imaging with arbitrarily oriented antennas within the data acquisition aperture is studied. The variations of antenna orientations change both the polarization of the radiated electromagnetic fields and spatial radiation patterns, which are assumed to be constant in traditional imaging algorithms. To overcome these effects, a linear inversion scheme is proposed by considering the exact radiation patterns of antennas. In terms of implementation, spatially varied radiation patterns are needed for arbitrarily orientated antennas. To efficiently acquire the radiation patterns, we suggest an interpolation-based method to compute the accurate Green's functions. Moreover, thanks to the frequency domain modeling, the proposed linear inversion approach also facilitates the signal fusion of multi-band GPR measurements for improved imaging, which is necessary for some practical applications. Finally, the imaging performance of the linear inversion approach is demonstrated through a numerical experiment.