ROBUST LONG RANGE TARGET DETECTION ALGORITHM FOR AIRBORNE APPLICATIONS

Abstract

Early detection of air threats like aircrafts and missiles is becoming more and more important for advancing attack distance and response speed of modern hi-tech , weapons in defence applications. These targets at longer ranges, typically 70 km or :- more, appear as point or small targets` in ,visible and infrared image sequences.- _> Because of the lack , of apriori information about- target dynamics and structural information such as shape, the detection t of such targets becomes extremely challenging: Recently °many algorithms for detection of paint and small targets for airborne applications have been reported in the ',literature. `It is found that most, of these ..:, algorithms are unable to - adapt their behaviour to perform robustly in the rapidly changing environment. In addition, the high computational . complexity`. of reported; detection algorithms requires large processing hardware and leaves a little -sco-pe`for real time . implementation. Hence; a novel and robust long range target detection algorithm based on Adaptive . Selective Double Structuring _ element Top-Hat Transform (Adapt-Set-DSTHT) and -maximum entropy criterion is proposed in this -dissertation report. -Experimental results -show high probability of detection and low false alarms even for highly clouded scenario:,; r a The proposed :robust long range target detection algorithm has also,- been -implemented on PowerPC hardware: AltiVec vector processing unit of Power-PC` G4 is used to speed up the processing and lead to real-time detection of targets