CORRELATOR (DSMAC) SCENE SELECTION AND SCENE ENHANCEMENT

CURRENT DSMAC SYSTEM USES SPATIAL-BINARY PATTERN TO PERFORM SCENE MATCHING. SUCCESS DEMANDS CERTAIN RELIABLE TERRAIN CHARACTERISTICS BE PRESENT AND EXTRACTABLE. ERGO, THE PRESENT OF CERTAIN OTHER TERRAIN CHARACTERISTICS MIGHT PRECLUDE SUCCESS, SUCH AS LINEAR FEATURES LIKE ROADS AND STREAMS. THIS IS THE SCENE AVAILABILITY PROBLEM. FURTHERMORE, FEATURES SELECTED FOR MATCHING ANALYSIS SHOULD BE IMMUNE TO SUDDEN ENVIRONMENT/WEATHER CHANGES; OTHERWISE, FEATURES IN THE REFERENCE SCENE MAY NOT EXIST IN THE SENSE SCENE. THIS IS THE FEATURE SET RELIABILITY PROBLEM. ACCORDINGLY, THESE TASKS ARE PROPOSED TO PROVIDE DSMAC WITH ADDITIONAL CAPABILITIES BY SOLVING THESE TWO PROBLEMS: (1) TO INVESTIGATE VARIOUS COMPETING TECHNIQUES TO ENHANCE DSMAC REFERENCE SCENE PREPARATION; (2) TO TEST COMPETING TECHNIQUES AGAINST SR & E'S ALGORITHMS BASED STABLE STRUCTURE THEORY; (3) TO INVESTIGATE ADVANCED ALGORITHMS TO EXTRACT LINEAR FEATURES TO ENHANCE DSMAC CAPABILITY; (4) TO COMPARE HUMAN EXPERT KNOWLEDGE AGAINST THE STABLE STRUCTURE THEORY AND ITS ALGORITHMS; (5) TO INVESTIGATE HARDWARE AND SOFTWARE INTEGRATION FOR DEVELOPING AN EXPERT SYSTEM. WITH SR & E EXPERIENCE IN THESE AREAS, AND THE FACT THAT THE MAJORITY OF THE ALGORITHMS HAVE BEEN DEVELOPED, THESE TASKS ARE ACHIEVABLE. FURTHERMORE, EXPERT KNOWLEDGE OF DSMAC WILL BE PROVIDED BY J. FELTENBERGER OF SAIC AS CONSULTANT TO SR & E.