A NOVEL FRACTAL-SENSITIVE IMAGE PROCESSOR FOR DETECTING GROUND TARGETS WITH SUPPRESSED SIGNATURES

WE PROPOSE TO DEMONSTRATE THE FEASIBILITY OF CONSTRUCTING AN ON-LINE IMAGE PROCESSOR WHICH WILL LOCATE ALL MANMADE COMPONENTS (TANKS, AIRCRAFT, ALL CURRENT CAMOUFLAGE, ETC.) IN A SCENE. THIS NOT ONLY GREATLY NARROWS THE REGIONS FOR SEARCH BUT ALSO MAKES THE RECOGNITION TASK MUCH EASIER, SINCE IT NO LONGER REQUIRES DISCRIMINATION AGAINST NATURAL BACKGROUND. THE SYSTEM IS BASED ON THE FRACTAL (STATISTICALLY SELF SIMILARITY AT ALL MAGNIFICATIONS) PROPERTIES OF NATURAL SCENES AND THE ABSENCE OF THAT PROPERTY IN MANMADE OBJECTS. SEVERAL COHERENT OPTICAL SPACE INVARIANT MONITORS OF FRACTAL DIMENSIONALITY WILL BE DEVELOPED AND TESTED. A DETAILED ENGINEERING DESIGN OF A COMPACT, ROBUST AND POWER EFFICIENT FRACTAL-SENSITIVE IMAGE PROCESSOR (FIP) WILL BE PRESENTED BASED ON THE EXPERIENCE OF DEMONSTRATING SUCH A SYSTEM. ANTICIPATED BENEFITS/POTENTIAL COMMERCIAL APPLICATIONS - THE PROPOSED FRACTAL-SENSITIVE IMAGE PROCESSOR IS APPLICABLE NOT ONLY TO DOD, BUT ALSO TO DEA, EPA AND OTHER GOVERNMENT ORGANIZATIONS THAT DEAL WITH CONCEALMENT. FOR THE DOD, THE ABILITY TO SEGREGATE NATURAL AND MANMADE COMPONENTS OF A SCENE IS OF VALUE IN EVERYTHING FROM RECONNAISSANCE TO TRACKING. POTENTIAL COMMERCIAL APPLICATIONS INCLUDE MEDICAL MONITORING, PROCESS CONTROL, ETC.