Improving mine recognition through processing and Dempster–Shafer fusion of ground-penetrating radar data

Abstract

A method for modeling and combination of measures extracted from a ground-penetrating radar (GPR) in terms of belief functions within the Dempster–Shafer framework is presented and illustrated on a real GPR data set. A starting point in the analysis is a preprocessed C-scan of a sand-lane containing some mines and false alarms. In order to improve the selection of regions of interest on such a preprocessed C-scan, a method for detecting suspected areas is developed, based on region analysis around the local maxima. Once the regions are selected, a detailed analysis of the chosen measures is performed for each of them. Two sets of measures are extracted and modeled in terms of belief functions. Finally, for every suspected region, masses assigned by each of the measures are combined, leading to a first guess on whether there is a mine or a non-dangerous object in the region. The region selection method improves detection, while the combination method results in significant improvements, especially in eliminating most of the false alarms.