Navy achieves real-time autonomous target detection
Researchers at the Naval Research Laboratory (NRL; Washington, DC) have demonstrated autonomous, real-time, hyperspectral detection of airborne and ground targets using real-time, autonomous cuing of a high-resolution imager. Targets were designated with pointing optics and a pulsed laser. The work, performed under the NRL "Dark Horse" program, demonstrated the capabilities needed for autonomous uninhabited combat air vehicles (UCAV). "This demonstration marks the first ste¥toward a realizable UCAV system," said Thomas Giallorenzi, head of the NRL Optical Sciences Division.
The airborne tests were conducted on an Orion (P-3) aircraft. Hyperspectral detection algorithms developed by NRL were implemented on a real-time processor. Two algorithms, operating simultaneously in an "AND" mode to reduce false alarms, produced a cuing signal upon target detection. Cued high-resolution imagery was obtained using a high-resolution reconnaissance framing camera containing 25 million pixels, and data were transmitted to a ground station at NRL. The capability to autonomously detect and cue on moving targets in real-time was demonstrated for ground targets as well as for a target aircraft flying 4000 feet below the P-3.