Combination of GPR and metal detector for mine detection

This paper reports the progress of a three-year project with the objective to demonstrate multisensor techniques for mine surveillance. Data fusion techniques are using ground penetrating radar (GPR) of type ERA and a metal detector (MD) with induction coil of type Schiebel. The integration with a biosensor has been postponed until further technical development. The data fusion process runs in three main steps. First all sensor reports are converted to likelihood's according to type. The likelihood's are based on a priori test measurements with buried objects of different materials and sizes (low metal content mine, metal shrapnel, wax, stone, sand). In the second step the joint sensor report during a given time period, e.g. one measurement with MD/GPR, is computed as a product of probabilities. Finally, nearby measurement points are updated along the scanning MD/GPR system using a Bayesian fusing method. Compared with single sensors, the two-sensor system results in enhanced probability of detecting an object with both metal content and a surrounding shell of contrasting material, e.g. a low-metal content anti-personnel mine. At the same time the false alarm rate is significantly lowered.