Object Recognition Results Using MSTAR Synthetic Aperture Radar Data

This paper outlines an appr oachand experimental results for Synthetic Aperture Radar (SAR) object recognition using the MSTAR data. With SAR scattering c enter locations and magnitudes as features, the invariance of these featur esis shown with obje ct articulation (e.g., rotation of a tank turr et) and with external con guration variants. This scatter er location and magnitude quasi-invariance is used as a basis for development of a SAR recognition system that successfully identi es articulated and non-standard conguration vehicles based on non-articulated, standard recognition models. The forced recognition results and pose accur acyare given. The e e ctof di er entconfusers on the receiver operating char acteristic(ROC) curves are illustrated along with ROC curves for conguration variants, articulations and small changes in depression angle. R esults ar e given that show that integrating the results of multiple recognizers can lead to signi cantly improved performance over the single best recognizer.