Detection of low-signature targets in rough surface terrain for forward-looking ground penetrating radar imaging

We develop an image-domain target detector for forward-looking ground penetrating radar (FLGPR) applications. An FLGPR offers the advantage of standoff sensing for detecting ground targets, but the target responses are more vulnerable to interference scattering arising from interface roughness and subsurface clutter. The proposed detection scheme draws all inferences regarding target and clutter statistics from the data measurements. More specifically, it iteratively adapts to the target and clutter statistics of the FLGPR images. Both single- and multi-aperture radar configurations are considered and the detection performance of each configuration is evaluated using electromagnetic modeling data.