Surface reflection removal in compressed sensing GPR and sparse subsurface imaging

Sparsity enhanced imaging methods for stepped-frequency continuos wave ground penetrating radars (SFCW GPRs) do not remove ground reflections. If the surface is rough and there are near surface targets, these targets can not be detected because of ground reflections. Standard methods can not be directly applicable for removing the ground reflections because of random and fewer number of measurements. In this work, a surface reflection removal method for random and fewer measurements is developed and compared with standard ground reflection methods. Simulation results show that the new method can remove ground surface reflections better than standard ground bounce methods and shallow targets can be detected with sparsity enhanced imaging methods.