Algebraic Algorithms Dramatically Enhance Surveillance Imagery

  • Published
  • By Maria Callier
  • Office of Scientific Research
ARLINGTON, Virginia --  Technology developed through Air Force Research Laboratory funding of two innovative mathematicians, Dr. Myoung An and Dr. Richard Tolimieri, is expected to advance object and target detection for the Air Force. The pair's work is founded in their remarkably insightful use of sophisticated algebraic theories called groups, rings, and fields. Leveraging these complex mathematical structures, the scientists produced novel algorithms that will promote the effective review of photographic, video, and radar images and, in turn, facilitate optimal military planning and order of battle.

A related technology, also crafted by Drs. An and Tolimieri but for US Navy purposes, earlier proved successful in detecting shallow-water mines by means of sonar (which operates much like radar but with sound waves instead of electromagnetic waves). The challenge inherent to this effort was to match the algebraic arrangements to the problems--and corresponding data index sets--at hand. Ultimately, the lab-sponsored adaptation of the established detection methodology will aid the AF in large-scale, computerized examination of surveillance imagery. Further, by introducing even more advanced mathematical structures, the researchers believe they will eventually realize enhancements that bolster the detection of subtle patterns or features under inclement conditions (e.g., dust, fog) and in the presence of foliage or comparable visual obstructions.

Prior to this technology discovery, careful human review of large amounts of surveillance material required a considerable time investment. While the new algorithms cut this substantial outlay by 99%, the human factor will remain essential in several procedural aspects, such as in the validation and verification of transmitter and receiver configurations. In addition to the technology's intended use in supporting the critical area of object/target detection, it could continue to evolve, perhaps one day benefiting the speech and language recognition applications of tomorrow's medical and linguistic fields, or otherwise proffering potentially lifesaving advances in various combat scenarios of the future.