Two Wavelet-based Algorithms for Chemical Recognition Using Transmission Terahertz Spectral Imaging Through Turbid Media

Abstract

Dielectric heterogeneity in addition to wavelength-scale air voids and granular particles lead to significant electromagnetic scattering in the transmission-mode THz-TDS. These scattering processes result in the distorted or obscured resonant signatures, a frequency-dependent extinction loss, and the appearance of anomalous spectral artifacts, imposing severe challenges on the real-world automatic material characterization schemes. In this work, we use a combination of the wavelet multiresolution analysis, the principal component analysis, and the bimodality coefficient to simultaneously identify all the resonant frequencies of a heterogeneous sample embedded beneath a highly scattering turbid medium.