Title :
Classification of aerospace targets using superresolution ISAR images
Author_Institution :
Dept. of Electr. & Electron. Eng., Pretoria Univ., South Africa
Abstract :
This paper describes investigations into a recognition system that takes the 2-D ISAR (inverse synthetic aperture radar) image of an aerospace target as input and classifies the target based on features calculated from the image. Four types of features were implemented, namely geometrical moments, invariant features based on moments, shape features, and quantized energy strips. Nearest-neighbour and neural-net classifiers are considered
Keywords :
aircraft; image classification; image resolution; neural nets; radar applications; radar imaging; radar target recognition; synthetic aperture radar; 2-D ISAR image; aerospace targets classification; aircraft; geometrical moments; image features; invariant features; inverse synthetic aperture radar; nearest-neighbour classifiers; neural-net classifiers; quantized energy strips; radar target recognition system; shape features; superresolution ISAR images; Aerospace engineering; Africa; Aircraft; Frequency measurement; Image resolution; Multiple signal classification; Neural networks; Signal resolution; Strips; Target recognition;
Conference_Titel :
Communications and Signal Processing, 1994. COMSIG-94., Proceedings of the 1994 IEEE South African Symposium on
Conference_Location :
Stellenbosch
Print_ISBN :
0-7803-1998-2
DOI :
10.1109/COMSIG.1994.512452
