ADVANCED TRACKER DEVELOPMENT USING FUZZY SET DETECTOR THE FOURIER MELLIN TRANSFORM AND AN ARTIFICIAL NEURAL NETWORK CLASSIFIER

ADVANCED TRACKING WITH ATTITUDE INFORMATION IS PROPOSED BY ANDRISANI AND KHUL (1985) FOR AIRPLANES. ORIENTATION EXTRACTION IN REAL TIME CAN IMPROVE GROUND VEHICLE TRACKING. WE PROPOSE USING TWO FILTERS TO ACCOMPLISH THE TASK: ONE TO PERFORM REAL TIME PHYSICAL TRACKING, AND THE OTHER TO PERFORM LEAD ANGLE PREDICTION. BOTH FILTERS ARE CONTROLLED BY ONE ATR SYSTEM WHICH PERFORMS TARGET DETECTION IN REAL TIME, AND ORIENTATION EXTRACTION AT A COARSER TIME INTERVAL. FOURIER BASED SHAPE MATCHING OFTEN YIELDS INCORRECT ORIENTATION FOR ITS ORIENTATION INVARIANT PROPERTY; HENCE THE FOURIER-MELLIN TRANSFORM IS PROPOSED AS AN ALTERNATIVE. IN ADDITION, THE SR&E NEURAL NETWORK WILL BE USED TO PERFORM SHAPE MATCHING BECAUSE IT IS NOT ORIENTATION INVARIANT. SINCE OUR TARGET DETECTOR IS A PIXEL BASED FUZZY SET ANALYZER, TRACKING CAN BE ACHIEVED IN REAL TIME. IN ADDITION, SINCE SEGMENTATION IS THE SUM OF DETECTED PIXELS, SHAPE MATCHING FOR BUILDING THE SECOND FILTER CAN BE EXPECTED IN REAL TIME AS WELL. THE FEASIBILITY OF THIS PROPOSED DUAL-FILTER TRACKING WILL BE TESTED WITH PHASE I EFFORT.