摘要
以棉涤、棉氨、粘涤、棉锦、羊腈、锦涤六种纺织布料为研究对象,采集近红外光谱数据,进行一阶求导和矢量归一化预处理后,采用标准算法计算样品间的光谱距离,并利用Ward氏算法对样品进行聚类分析.试验对三种、四种、五种布料分别进行了聚类,所有不同种类的样品都得到了正确的分类,验证了近红外光谱检测法应用于纺织品聚类分析中的可行性。通过对聚类算法的分析,提出了局部回归算法在纺织品聚类分析中的应用,为进一步研究近红外光谱技术在纺织布料聚类分析中的应用提供了建议。
Six kinds of textile samples including cotton-polyester, cotton-polyurethane, viscosepolyester, cotton-polyamide, wool-polyacrylonitrile and polyamide-polyester were selected as the research objects. The near infrared spectral data were collected from the samples and were preprocessed through first derivation or standard normalization. A standard algorithm was used to calculate the spectral distance among the samples and the Ward＇s algorithm was used to make clustering analysis on the samples. All of the samples were classified correctly. Thus, the feasibility of the application of near infrared spectroscopy （NIR） in the clustering analysis of textile was verified. A new method, local weighing agression which was useful for the further research on the application of NIR in clustering analysis of textile was proposed.
出处
《红外》
CAS
2009年第1期31-35,共5页
Infrared
作者简介
柴金朝（1982-），男，浙江杭州人，中国计量学院光学工程专业在读研究生，主要研究方向为服装面料成分及含量的近红外光谱检测。E-mail：chaijinchao@163．com
