Spectral characterization and discrimination of synthetic fibers with near-infrared hyperspectral imaging system.

Abstract

Synthetic fibers account for about half of all fiber usage, with applications in every textile field. The phenomenon of fiber composition not matching the label harms consumer interests and impedes market development. Hyperspectral imaging technology as a potential technique can be utilized to discriminate mass synthetic fibers rapidly and nondestructively and achieves the functions that traditional Fourier transform infrared instruments do not have. On the basis of investigating the impact of dope-dyeing and wrapping processes on spectra, the spectral features (from 900 to 2500nm) of six categories of synthetic fibers were extracted with a hyperspectral imaging system. A principal component analysis-linear discriminant analysis model was developed to discriminate the chemical content of fibers in different colors and structures, which showed 100% discrimination accuracy. Results demonstrated the feasibility of a hyperspectral imaging system in synthetic fiber discrimination. Therefore, this method offers a more convenient alternative for textile industry on-site discrimination.