Abstract
Nonmetallic mine detection with GPR is challenging because of poor target/soil contrast and clutter from the random rough ground surface. Greater soil moisture implies higher dielectric constant, which in turn increases contrast and scattering of radar waves, but also increases the contrast between the air and the soil, increasing clutter. The finite difference time domain method for modeling impulse radar is use to quantify the effects of clutter on the target signal for both ideal smooth ground surfaces and rough surfaces for uniformly dry, wet, and a profiled moisture distribution characteristic of short-time watering. Results indicate that each uniform moisture case offers detection features that allow for the distinct temporal and spatial separation of target and clutter signals, but that non-uniform soil moisture makes nonmetallic mine detection problematic