The Effects of Geometric Scattered Signal Waveform Modeling on Target Identification Performance

Abstract

Target identification from scattered signals using time domain techniques depend significantly on the waveform. Recently a novel feature set is proposed which encounter structural properties of the waveform and collects local extrema points to model the scattered signal via triangularization. Then using this piecewise model it extracts several morphological features and employs them for target identification through classification. This study expands that approach by modeling the scattered signal with other geometric shapes and accordingly by enriching the feature set. Such an approach requires careful representation of the waveform model since more than one morphology is considered to represent sub-waves of the waveform. The effects of the proposed approach are observed by applications on spherical targets having different size and material type.