Persistent IR Detection and Tracking

Recent employment of airborne persistent surveillance systems by the Department of Defense has provided a significant force multiplier to military operations. These systems are capable of providing real-time imagery from airborne sensor systems to a ground station in order to provide situational awareness to the warfighter. However, limited ground sample distance results in dismount targets that are subpixel or no more than 3-by-3 pixels in size, leading to difficulties with both image registration and noise suppression. Furthermore, low contrast images are typically encountered making it difficult to detect targets in scenes with clutter. Numerica''s expertise in multi-target tracking and target detection will enable the development of a software suite capable of real-time automatic detection and tracking of dismounts in MWIR imagery. Track-before-detect methods will be necessary to manage the low contrast imagery with point sized targets. In addition, robust tracking methods designed to handle multiple targets will be required. Specifically, multiple hypothesis tracking (MHT) provides deferred decision-making for the data association problem that is central to multi-target tracking and automated track initiation. The primary focus of this proposed research is to develop and implement a multi-target tracking system incorporating innovative track-before-detect methods tightly integrated with a mature MHT. BENEFIT: The developmental sections of the software developed under this effort can be readily translated into a real-time system for direct use by the military and civilian reconnaissance communities. Specific civil application areas include border patrol, urban surveillance and search and rescue. The solutions to these problems are of direct interest to the U.S. armed forces as well as the Department of Homeland Security. The recent incorporation of large format MWIR sensors on airborne platforms will be appropriate for homeland defense applications that require semi-autonomous sensor coverage along with high-confidence tracking solutions. Our ongoing relationship with Lockheed Martin and Boeing on efforts related to tracking system development will provide an avenue for commercialization. It is anticipated that the algorithms developed and results obtained by Numerica under this SBIR will be incorporated by in their continued enhancements of defense systems. Direct support of this goal is being provided by Lockheed Martin through their offer to provide Numerica with access, free of charge, to the Lockheed Martin MS2 Technology Collaboration Centers. Numerica and Boeing are teamed with Ball Aerospace to support the National Air and Space Intelligence Center (NASIC) for the exploitation and development of technology that utilizes persistent sensor data.