近红外光谱在快速检测棉制品中含棉量的应用.pdfVIP

28 3 ( ) Vo.l 28 No. 3 2007 5 Journal of Jiangsu University(NaturalScience Edition) ay 2007 1 1 2 1 陈 斌 , 崔 广, 金尚忠, 王小天 ( 1. , 212013; 2. , 310000) : 以两组棉制品为研究对象,利用遗传算法提取有效 红外光谱信息,采用偏最小二乘法( par- tial leastsquares, 简称 PLS)建立了棉制品中含棉量的 红外定量的校正模型,讨论了遗传算法提取 有用信息的具体应用方法.结果表明:棉制样品的 红外光谱经过遗传算法提取有用信息 (波长优 选)后, 通过 一定数据预处理方法,运用 PLS建立的定量校正模型, 可以大大降低数据运算量,为 红外光谱分析技术应用于棉制品中含棉量的预测, 提供了理论依据和实用方法. : 棉制品; 红外光谱; 遗传算法; 偏最小二乘法; 校正模型; 数据预处理 : O657. 33 : A : 1671- 7775( 2007) 03- 0185- 04 Application of near infrared spectra in rapid inspection of cotton contents 1 1 2 1 CHEN B in , CUI Guang , JIN Shang-zhong , WAN G X iao-tian ( 1. School of Food and Biological Engineering, Jiangsu University, Zhenjiang, Jiangsu 212013, China; 2. School of Information Engineering, China JiliangUniversity, H angzhou, Zhejiang 310000, China) Abstract: Two groups of cotton fabricswere taken as research objects. The usefulnear-infrared spectrum information was picked up by using genetic algorithms, and the near-infrared quantitative correction modelwas established by partial least squares ( PLS) method. The particular application methods of available information picked up by genetic algorithmswere discussed. Itwas proved that the quantitative correction model, whichwas established by definite data pretreatment and PLS, after the near-infrared spectrum of cotton fabric sampleswere distilledby geneticalgorithms, may greatly reduce thedata operation quantity.