Mine detection under rough ground surfaces using 2-D FDTD modeling and hypothesis testing

Summary form only given. This paper attempts to investigate hypothesis test performance in mine detection, based on numerically simulating the wideband scattering of experimentally measured ground penetrating radar (GPR) signals by realistic, dispersive soil interfaces. We employ a 2D finite difference time domain (FDTD) method to analyze the delay and amplitude characteristics of ground-scattered waves as a function of roughness parameters. In addition, we apply binary hypothesis tests to the signals obtained using physics-based signal processing techniques to investigate the presence of the target at certain depth. We quantify the detection performance in terms of the spatial distribution of transmitter and receiver.