Recognition System in SAIP

VOLUME 10, NUMBER 2, 1997 THE LINCOLN LABORATORY JOURNAL 187 T   automatic target recognition (ATR) is to detect and recognize objects, such as tanks, in images produced by a laser radar, a synthetic-aperture radar (SAR), or an infrared or video camera. Current ATR methods are not fully automatic, relying instead on a partnership between computer processing and human analysis. Although computer algorithms accomplish much preliminary ATR work of sifting through millions of data pixels per image to identify candidate targets, image analysts play a crucial role in the recognition process by reviewing the ATR algorithm results. In a typical conflict situation, seven image analysts and one supervisor perform image interpretation for about thirty minutes to produce a target report that details the presence of missile launchers and the ground order of battle vehicles such as tanks, armored personnel carriers, and trucks. The advent of low-cost, wide-area sensor platforms on unmanned air vehicles (UAV) will generate a profusion of data that dramatically increases the demands on image analysts [1]. For example, Teledyne Ryan Aeronautical’s Global Hawk UAV is designed to provide sustained high-altitude surveillance and reconnaissance at large standoff ranges. Its mission goals include an operating range of 3000 nautical The Automatic TargetRecognition System in SAIP