A multimodal Matching Pursuits Dissimilarity Measure applied to landmine/clutter discrimination

The Matching Pursuits Dissimilarity Measure (MPDM) is an effective way to to compare signals that are sparsely approximated using a Matching Pursuits method. The CAMP algorithm uses an MPDM distance measure in Competitive Agglomeration clustering to model and classify signals. The MPDM approach can only compare signals originating from a single source. Many landmine detection systems use multiple sensors to make simultaneous measurements of the same region of interest. In this paper we propose a Multimodal MPDM that can be used with CAMP to fuse signals from multiple sensors. We demonstrate the effectiveness of the Multimodal MPDM over the single sensor MPDM in improving discrimination of landmines from clutter objects.