분광법과 케모메트릭스 방법 연계를 통한 다른 옷감 섬유들의 일치성 판별

Abstract

The identification of textile fibers with different colors using visible and Raman spectroscopy has been studied in this thesis. For this purpose, both visible and Raman spectra of textile fibers with 6 different colors (black, green, pink, red, violet and yellow) were initially collected using a microscope. Then, hit quality index (HQI) indicating the degree of spectral matching between two separate spectra was employed for the identification. When visible spectroscopy was solely used, the accuracy in the measurement of black fibers was considerable poor and the accuracies in Raman measurements of pink and yellow fibers were unsatisfactory. Since both spectroscopic methods were able to provide spectral information with different characteristics, their combination would be cooperatively synergetic in the identification. For the investigation, two HQIs obtained from both methods were simply averaged and then the average HQIs were used for the identification. The combined use of both data improved the identification accuracy in all cases. The poor accuracy in the visible spectroscopic measurement of black fibers was strongly backed up by 3 simultaneously employing the Raman spectroscopic data. Vice versa, the unsatisfactory accuracies in Raman measurements of pink and yellow fibers were greatly improved by incorporating the visible spectroscopic data. The resulting accuracies ranged from 58.3 to 100%, clearly elevated in comparison with those obtained using a single spectroscopic method. HQI was simple and straightforward to determine degree of spectral matching without blackbox-like modeling procedure in case of employing complex chemometric methods. For final field application, the number of samples should be substantially increased to secure a wide spectral database and the further inclusion of such as infrared (IR) and near-infrared (NIR) spectroscopy would be worthwhile.