Hierarchical Reconstruction and Structural Waveform Analysis for Target Classification

Abstract

Classification of objects from scattered electromagnetic waves is a difficult problem, as it heavily depends on aspect angle. To minimize this dependency, distinguishable features can be used. In this paper, we propose a target identification method in the resonance scattering region using a novel structural feature set based on the scattered signal waveform. To obtain robustness at low signal-to-noise ratio (SNR), a multiscale approximation is used for distortion correction prior to the feature extraction. This is achieved by an overlapping grid hierarchical radial basis function (HRBF $_{\mathrm {\mathrm {OG}}})$ network topology, which is demonstrated to outperform existing HRBF techniques. The results obtained from the simulations and the measurements performed for various targets show high accuracy for classification with the proposed feature set, robustness through the use of HRBF at low SNR, and efficient computation in real time.