MAP image reconstruction for land mine and IED detection using ground penetrating radar

In this paper, we introduce an iterative maximum a posteriori (MAP) algorithm that reconstructs reflectance images from ground penetrating radar (GPR) data. Typically, the scene of interest (SOI) being illuminated by the GPR contains only a few electromagnetic wave scatterers. We account for this prior information through the use of a novel sparsity-promoting density function, which we call the Butterworth prior. The Butterworth prior can mimic the Dirac Delta function or a uniform density function, and possesses the added benefit of mathematical tractability. We have tested our algorithm on real data provided by the US Army Research Laboratory.