Key points are not available for this paper at this time.
An efficient technique is developed to recognize target type using one-dimensional range profiles. The proposed technique utilizes the Multiple Signal Classification algorithm to generate superresolved range profiles. Their central moments are calculated to provide translation-invariant and level-invariant feature vectors. Next, the computed central moments are mapped into values between zero and unity, followed by a principal component analysis to eliminate the redundancy of feature vectors. The obtained features are classified based on the Bayes classifier, which is one of the statistical classifiers. Recognition results using five different aircraft models measured at compact range are presented to assess the effectiveness of the proposed technique, and they are compared with those of the conventional range profiles obtained by inverse fast Fourier transform.
Kim et al. (2002) studied this question.