Explosives identification model in reflection mode for THz security system

The aim of this paper was to obtain the identification model in refection mode. Results Time Domain Spectroscopy were used to prepare our algorithm. This study has focused on developing several feature extraction methods with intuitive justifications in the problem space. A related problem to feature extraction is that of feature selection. For this reasons this extraction and selection methods of THz spectra are introduced. Then a complete THz classification framework including feature extraction scheme and Mahalanobis classifier was presented. Our results confirm the possibility of application of the model in real THz stand-off security system.