Broadband bistatic coherent and incoherent detection of buried objects beneath randomly rough surfaces

This paper investigates the application of angular correlation function (ACF) processing for target detection when a metallic object beneath a randomly rough soil surface is sought by ground penetrating radar (GPR). ACF processing is an inherently coherent and multistatic approach. The author compares its results to those from candidate multistatic incoherent processing strategies. In 2D numerical simulations, the author assumes that the antennas are elevated above the soil surface, and that the only significant source of environmental randomness is the soil surface roughness. To expand the basis for the ACF, averaging is performed over both frequency and limited spatial shift. The author takes pains to use realistic GPR problem parameters, together with relevant and varied target geometries. Results are evaluated over an ensemble of geometrical realizations to gain some statistical picture of each method's behavior and performance. Under the parameter limitations applied, combined frequency and spatial averaging are required to produce recognizable ACF behavior. Contrary to expectations, the ACF processing performs best in the vicinity of the "memory line," not when it is avoided.