Electromagnetic Target Classification of Small-Scale Aircraft Modeled by Conducting Wire Structures Using a Natural Resonance Based Feature Extraction Technique
Mehmet Okan Ersoy, Gönül Turhan-Sayan*
Electrical and Electronics Eng. Dept., Middle East Technical University, Ankara, Turkey
The problem studied in this paper, is the design of an electromagnetic target classifier for small-scale aircraft targets by using a natural resonance based feature extraction technique supported by feature fusion. The aircraft targets are modeled by perfectly conducting straight thin wire structures and the electromagnetic fields back-scattered from targets are numerically generated. The design technique used in this work was formerly proposed to minimize the aspect dependency of extracted target features to achieve high recognition accuracy, fast real-time decision while using a minimal reference database of scattered fields. This technique, which uses the Wigner-Ville Distribution (WD) and the principal component analysis (PCA) as the main signal processing tools, has been applied to dielectric spherical targets with success [1]. In the present paper, the technique is applied to perfectly conducting targets of moderately complicated geometries for the first time. The resulting classifier is successfully demonstrated to recognize five wire-modeled aircraft targets based on a reference database that is constructed at only three different aspect angles.
IEEE Antennas and Propagation Society International Symposium
2005
June
4A
468-471
Due to copyright restrictions articles are not available for download. They are however indexed on this site and search results will therefore correctly indicate if such terms are present in the articles.
