Classification of aerospace targets using superresolution ISAR images

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.