Automated Ship and Small Craft Classification Tools for ISAR Imagery

MARK Resources proposes to develop ISAR-based classification techniques for ships and small craft, based on extracting 3-D scatterer positions of major ship scatterers. The proposed extraction procedure leverages prior ONR-funded work that allows us to select imaging intervals that give topviews and sideviews. We propose to associate the scatterer positions in the two views by analyzing distortions of a topview as the imaging interval is shifted slightly in time. The measurements will be based on Complex-Image Analysis, a radar resolution theory and processing technology appropriate for highly resolved man-made targets, which achieves the inherent resolution capability of radar. With Complex-Image Analysis, the positions of major scatterers persist over large changes in target orientation, so require only a small comparison database, which is easily obtained. The extracted positions are robust against contamination by interfering signals, insensitive to changes in configuration, and allow classification of previously unknown targets. The classification algorithms and software we propose to develop are compatible with existing airborne and ship-borne radars, and are computationally efficient so that the classification capability could be implemented on a UAV.