Early study of Near Infrared Spectroscopy(NIR)for different Textiles Identification and Quantification,Part Ⅰ
Many new textile materials have been introduced to the market in recent years.Most new materials are not made of only one type of fibers but more commonly of mixtures.To identify different fibers and the percentage of each fiber in the textile are very important.However,most textile industries and relevant inspection bureaus in the developing countries are still using classical chemical analysis standard methods to identify and quantify various fibers mixed in the textiles.Near Infrared Spectroscopy has been utilized for feasibility tests to identify various textile materials.A NIRSystem 6500 instrument was used in the study.Over 100 samples were scanned,spectra wavelength range is from 1100 nm to 2500 nm.These raw materials were pure cotton,polyester,lycra,ecylon,wool,ramie,linen,polyamide,silk,linen,viscose,etc.;as well as the textiles of mixed fiber of 2-4 different materials.The chemical structures were considered in order to interpret specific NIR absorption bands.By using Chemometrics tools such as second derivatives,partial least squares regression(PLS),principal component analysis(PCA),a few specific SIMICA(PCA)models are made to differentiate cotton,polyester and polyamide.The final validation of these models was made by comparing their prediction results with the wet chemistry analysis(used as reference methods)results.The general Correlation Coefficients of calibration models for most materials are above 95 percent and are with RMSEP within 2.0.
NIR spectroscopy PLS SIMICA PCA regression textile fiber identification qualitative and quantitative model
W.G.HANSEN Y.Y.LIAO
NIR application Consultants Inc.Katy TX 77494,USA Jiangxi Entry-Exit Inspection and Quarantine Bureau,Nan Chang,Jiangxi,P.R.China
国内会议
桂林
英文
155-162
2012-09-12（万方平台首次上网日期，不代表论文的发表时间）