MOST EXISTING DETECTION ALGORITHMS ARE "SINGLE FRAME" TECHNIQUES: THE IR OR RADAR SENSOR PROVIDES A SINGLE IMAGE OF THE AREA UNDER SURVEILLANCE; THIS IMAGE IS THEN PROCESSED TO DETECT THE PRESENCE OF TARGETS.

MOST EXISTING DETECTION ALGORITHMS ARE "SINGLE FRAME" TECHNIQUES: THE IR OR RADAR SENSOR PROVIDES A SINGLE IMAGE OF THE AREA UNDER SURVEILLANCE; THIS IMAGE IS THEN PROCESSED TO DETECT THE PRESENCE OF TARGETS. THESE TECHNIQUES OFTEN FAIL TO PERFORM ADEQUATELY IN A DENSE TARGET ENVIRONMENT. IN THIS PROPOSAL WE OUTLINE A DIFFERENT APPROACH BASED ON USING MULTIPLE FRAMES TO PERFORM DETECTION. IN PHASE I WE WILL STUDY THE PERFORMANCE IMPROVEMENT ACHIEVABLE BY THE MULTI-FRAME DETECTOR, COMPARED TO MORE CONVENTIONAL DETECTION TECHNIQUES, AND DEMONSTRATE ITS FEASIBILITY. IN PHASE II WE WILL DEVELOP PRACTICAL IMPLEMENTATIONS OF THE MULTI-FRAME DETECTOR.