纺织纤维的快速分析对纺织品生产、质量监控、贸易和市场监督具有重要意义。本研究采用NIRFlex N-500型傅立叶近红外光谱仪在4 000~10 000 cm-1光谱范围内采集棉、麻样品的漫反射光谱,根据棉、麻纤维的取向度存在显著差异这一特征,利用无信息变量消除法(UVE)、连续投影算法(SPA)以及CARS(competitive adaptive reweighted sampling)方法对棉、麻样品的取向度指标的近红外光谱进行变量优选,确定取向度在该区域的特征波段,在此基础上,应用PLS(partial least squares)建立棉、麻纤维取向度定量分析模型,利用该模型可对棉、麻样品的取向度进行预测,根据取向度大小进行棉、麻的鉴别。该研究结果验证了依据取向度进行棉、麻近红外光谱区间差异信息波段选择的正确性和利用近红外技术对棉、麻纤维快速判定的可行性。 Textile fiber rapid analysis of textile production, quality control, trade and market supervision is of great significance. In this study, the NIRFlex N-500 FT-NIR spectrometer was used to collect the diffuse reflectance spectra of cotton and hemp samples in the spectral range of 4 000 to 10 000 cm-1. According to the significant difference in the orientation of cotton and hemp fiber, UVE, SPA, and CARS (competitive adaptive reweighted sampling) methods were used to optimize the near-infrared spectra of the orientation indices of cotton and hemp samples to determine the orientation degree in the characteristic band Based on this, a PLS (partial least squares) model was established to quantitatively analyze the degree of orientation of cotton and hemp fiber. Using this model, the degree of orientation of cotton and hemp samples can be predicted, and the identification of cotton and hemp can be made according to the degree of orientation. The results of this study validate the correctness of band selection of the difference information between cotton and hemp near infrared spectra according to the degree of orientation and the feasibility of rapid determination of cotton and hemp fiber by using near infrared technology.