Fiber-optic spectroscopic sensor for reactive dye mixture spectrum synthesis in textile industry

A method for the automatic synthesis of the Ultraviolet-Visible-Near Infrared (UV-Vis-NIR) absorption and transmission spectra of dye mixtures based on the absorption characteristics of their individual dyes is proposed in this paper. Multiple Linear Regression models (MLR) of each dye are obtained using a fiber-optic set-up operating in the 200-1100 nm wavelength range. Textile dyes are thick and dense and, consequently, optically opaque. This gives rise to high absorbance values which do not permit the direct comparison of different textile dye spectra for quality assurance purposes. The proposed multivariate method allows to construct a general model adjusted to the number and concentration of the dyes in the mixture. Furthermore, this dye spectrum synthesis can provide the optimum dilution factor needed to compare two different dye mixtures and test their similarity degree. With the proposed system, the obtained spectral correlation coefficients between the measured and synthesized spectra of a dye mixture are greater than 99% for both the transmission and absorption spectra.