苎麻纤维素含量近红外预测模型的建立与探讨  

Study on Building Prediction Model with NIRS to Test Ramie Cellulose Rapidly

作　　者：肖爱平[1] 冷鹃[1] 杨喜爱[1] 程毅[1] 廖丽萍[1] 田小兰[1] 聂晴岚[1] 

机构地区：[1]中国农业科学院麻类研究所,湖南长沙410205

出　　处：《中国麻业科学》2011年第4期189-191,213,共4页Plant Fiber Sciences in China

基　　金：公益性农业行业专项:苎麻;亚麻;黄红麻的高效生产与收获技术研究合同号nyhyzx07-018-11

摘　　要：本文旨在探讨用近红外光谱技术建立预测模型快速测定苎麻中纤维素含量的可行性。以80个样品组成校正集采用偏最小二乘法(PLS)建立近红外光谱信息与纤维素含量之间的定量预测模型,该模型的相关系数为0.9726。利用该模型对22个样品的纤维素进行预测,预测值与化学测定值之间的平均相对误差为1.17%。结果表明,利用近红外光谱技术建立预测模型可以快速测定苎麻纤维素含量,对苎麻育种和优质资源的筛选利用具有重大意义。This paper described the feasibility of building a prediction model with Near Infrared Spectra（NIRS） to test ramie cellulose rapidly.80 calibrating samples were tested to get the NIRS information and cellulose content respectively,then this model was built by using Partial Least Squares Method（PLS）,the coefficient of correlation（R） of the model is 0.9726.The average relative deviation（ARD） was 1.17% between the predicted values from the model and the tested values from the chemical method for 20 samples.The results showed that it was feasible to build a prediction model with NIRS to predict ramie cellulose content rapidly,which will be prosperous for breeding and selecting high-quality ramie varieties.

关 键 词：近红外光谱 苎麻 纤维素 

分 类 号：S563.1[农业科学—作物学]