Realistic modelling of ground penetrating radar for landmine detection using FDTD

A finite-difference time-domain (FDTD) algorithm is used to model and study the performance of ground penetrating radar (GPR) for anti-personnel (AP) landmine detection. A novel algorithm is proposed which creates the geometry of the vegetation for both grass and roots. Soil´s inhomogeneities as well as the rough surface are simulated using fractal correlated noise. Debye functions are used in order to simulate the frequency dependent dielectric properties of both the soil and of the vegetation. The antenna unit that has been employed in the model is based on a previously developed detailed antenna model approximating a well known commercial GPR antenna, and the target is the anti-personnel (AP) landmine PMA-1. Surface water puddles have been included into the models and their effects on the performance of GPR are investigated. Simulation results are realistic and provide a useful testbed for evaluating GPR processing approaches for landline detection.