Automated Ship and Small Craft Classification Tools for ISAR Imagery Navy SBIR FY2006.2 Sol No.: Navy SBIR FY2006.2 Topic No.: N06-122 Topic Title: Automated Ship and Small Craft Classification Tools for ISAR Imagery Proposal No.: N062-122-0720 Firm: Lambda Science, Inc. P.O. Box 238 Wayne, Pennsylvania 19087-0238 Contact: Joseph Teti Phone: (610) 581-7940 Web Site: www.lamsci.com Abstract: Automated ship and small craft classification tools for ISAR imagery is a challenging problem, but the use of advanced time-frequency processing techniques in conjunction with proven methods from statistical pattern recognition hold considerable promise for realizing a sound capability. LSI's approach leverages our state-of-the-art knowledge and expertise of the three primary aspects of the solution space: (i) ISAR image formation and time-frequency information extraction for feature vector formation, (ii) correct physical interpretation of feature vector components that include target and clutter scattering phenomena and kinematics attributes, and (iii) multi-dimensional Bayesian engine based decision theory to implement a statistically optimum ATR capability Benefits: The proposed effort will result in software tools that automatically classify ship and small craft from ISAR mode data. The capabilities of the ISAR ATR software can be tailored to easily support the Navy's acquisition needs involving legacy and future radar sensors (e.g., APS-137, APS-147, ISS and radar sensors for the MMA and BAMS platforms). The proposed effort will also provide direct benefit to the use of radar in the US Coast Guard's Maritime Domain Awareness (MDA) charter in support of meeting the needs of the Department of Homeland Security (DHS). Return