SMART MAP-TO-IMAGE MATCHING SYSTEM DEVELOPMENT

WE PROPOSE TO STUDY THE FEASIBILITY, AND SUBSEQUENTLY BUILD A PROTOTYPE, OF A SYSTEM FOR AUTOMATED MATCHING OF IMAGES AND MAP DATA BASES. THIS WILL BE BASED ON AN EVALUATION OF APPROPRIATE MAP DATA BASE STRUCTURES TO ADDRESS, SEARCH FOR AND RANK OBJECTS THAT ARE CANDIDATE GROUND CONTROL. IT WILL FURTHER BE BASED ON USING NOT ONLY POINT-TYPE CONTROL, BUT ALSO LINEAR AND AREAL FEATURES FROM THE DATA BASE. THE SENSING GEOMETRY NEEDS TO BE PHOTOGRAMMETRICALLY MODELED TO RELATE MAP OBJECTS TO CONJUGATE IMAGE FEATURES. THE ACTUAL MATCHING PROCESS IS ANTICIPATED TO EMPLOY THE CONCEPT OF "MAP-GUIDED" FEATURE RECOGNITION, WHERE MAP-GUIDANCE IS IN THE FORM OF DEFINED OBJECTS AND LIMITED SEARCH SPACES. PHASE I EFFORTS WILL CONCENTRATE ON AN ELABORATE FEASIBILITY STUDY WITH ALGORITHMIC EVALUATIONS AND COMPUTATIONS WITH TEST DATA SETS.