Surface penetrating radar image quality

The detection of buried or surface laid mines by ultrawideband, ground-probing radar systems is currently receiving much attention. The radar signature of the mine may be known, in which case a range of techniques can be used to characterise the received signal. These range from, in the case of surface laid mines, characterisation of the sequence of reflections caused by external reflections, creeping waves and internal reflections such as the Glory wave; singularity expansion methods (SEM); resonance characterisation via such techniques as Prony; through to wavelet analysis. For buried mines the range of classification options is more limited as the host material significantly modifies the impulse response of the target, but shallow mines may still be identifiable. The extent of validation trials on most of these processing methods is limited. This paper considers some of the parameters, whereby the discrimination of the spatial signature of the mine is based solely on a statistical assessment of the radar image data and a characterisation of the radar sensor.