In this paper we present the sensor-fusion results based on the measurements obtained within the European research project GEODE (Ground Explosive Ordnance DEtection system) that strives for the realisation of a vehicle-mounted, multisensor, anti-personnel land-mine detection system for humanitarian demining. The applied decision-level sensor-fusion techniques are Bayesian approaches, application of Dempster- Shafer theory fuzzy probabilities, rules, and voting techniques. For the evaluation of the performance of sensor fusion, we introduce a novel algorithm that provides a less biased estimate of the performance measured in probability of detection and probability of false alarm. The evaluation method differs from common performance evaluation methods in the sense that it takes into account the number of false alarms as well as the area of false alarms. Furthermore, application of this measure of false alarms leads to intuitive receiver-operating characteristic curves.