Clutter reduction and detection of landmine objects in ground penetrating radar data using likelihood method

In this paper, an algorithm is proposed to reduce clutter signal due to the reflection from the ground surface in ground penetrating radar (GPR) measurements. The proposed technique has been applied to a GPR, which has been used to detect subsurface anti-personnel (AP) landmines. A very simple model is used to model the GPR clutter and the target signal. This model has been used to derive a detector based on a likelihood ratio test using some estimated parameters. A new threshold formula has been derived and has been applied to the data. The data sources include 1 GHz pulse GPR data and 1 GHz to 4 GHz stepped- frequency GPR data, from laboratory measurements. In all measurements, the reflections from the surface have been removed or at least reduced in great quantities.