Approach to object classification using dispersive scattering centres

The dispersive scattering centre (DSC) model characterises high-frequency backscatter from radar objects as a finite sum of localised scattering geometries distributed in range. These geometries, along with their locations, can be conveniently used as features in a one-dimensional automatic object recognition algorithm. The DSC model´s type and range parameters correspond to geometry and distance features according to the geometric theory of diffraction (GTD). To demonstrate the viability of feature extraction based on the DSC model´s range and type parameters, a typical object classification experiment was performed. The experimental data contained direct range radar measurements of four model fighter aircraft of similar size and shape at 0° elevation and 0°-30° azimuth. After implementing DSC model feature extraction on these data, a fully-connected two-layer neural net obtained over 98% classification accuracy. In addition, DSC model feature extraction gave an approximately 85% reduction in the number of required features when compared to the numerous range bin magnitudes used in template matching techniques.