湖北农业科学 ›› 2021, Vol. 60 ›› Issue (12): 128-130.doi: 10.14088/j.cnki.issn0439-8114.2021.12.025

• 检测分析 • 上一篇    下一篇

不同产地马铃薯主要成分的近红外无损测定模型

吴延勇1,2, 史艳财3, 吕雪娟4, 丘立杭1,2, 罗含敏1,2   

  1. 1.广西壮族自治区农业科学院甘蔗研究所,南宁 530007;
    2.农业农村部广西甘蔗生物技术与遗传改良重点实验室/广西甘蔗遗传改良重点实验室,南宁 530007;
    3.广西壮族自治区中国科学院广西植物研究所,广西 桂林 541006;
    4.华南农业大学测试中心,广州 510642
  • 收稿日期:2021-03-17 出版日期:2021-06-25 发布日期:2021-07-19
  • 通讯作者: 罗含敏,主要从事植物生物技术研究,(电子信箱)lhmin26@163.com。
  • 作者简介:吴延勇(1972-),男,广西容县人,主要从事植物生物技术研究,(电话)13978700518。
  • 基金资助:
    广西农业科学院青年科学基金项目(2013YQ15)

Nondestructive testing model for main components of potato from different places by near infrared spectroscopy

WU Yan-yong1,2, SHI Yan-cai3, LYU Xue-juan4, QIU Li-hang1,2, LUO Han-min1,2   

  1. 1. Sugarcane Research Institute,Guangxi Academy of Agricultural Sciences,Nanning 530007,China;
    2. Key Laboratory of Sugarcane Biotechnology and Genetic Improvement,Ministry of Agriculture and Rural Affairs/Guangxi Key Laboratory of Sugarcane Genetic Improvement,Nanning 530007,China;
    3. Guangxi institute of Botany,Guangxi Zhuang Autonomous Region and Chinese Academy of Sciences,Guilin 541006,Guangxi,China;
    4. Instrumental Analysis Center of South China Agricultural University,Guangzhou 510642, China
  • Received:2021-03-17 Online:2021-06-25 Published:2021-07-19

摘要: 收集全国6个省(区)不同品种和分级的36份马铃薯(Solanum tuberosum L.)样品,以精确化学方法测定的成分含量值为依据,利用偏最小二乘回归法(PLS)建立马铃薯水分、还原糖和淀粉含量的近红外无损测定模型。结果表明,建立的马铃薯水分、还原糖和淀粉模型的决定系数(R2)分别为0.88、0.84和0.92,交互验证均方根误差(RMSCV)均小于1,验证集标准偏差与预测标准偏差的比值(RPD)均大于2.5。通过外部验证发现,模型预测的各成分含量决定系数R2均大于0.9,预测值与实测值的相对偏差较小。

关键词: 马铃薯(Solanum tuberosum L.), 水分, 还原糖, 淀粉, 近红外光谱

Abstract: 36 potato(Solanum tuberosum L.) samples involved in different species and grades were collected from 6 provinces in China. Based on the values determined by chemical methods, near infrared spectroscopy (NIRS) nondestructive testing model for the water, reducing sugar and starch content were constructed by partial least squares regression method, respectively. The results showed that the determination coefficients (R2) of constructed NIR model were 0.88 for water, 0.84 for reducing sugar, and 0.92 for starch. The root mean square error of cross validation (RMSECV) was less than 1, and the ratio of the standard deviation of the verification set to the predicted standard deviation (RPD) was all more than 2.5 in each NIR model. In addition, the results of external validation further showed that the R2 for three components which was predicted by each constructed NIR model was more than 0.9, and the relative deviation between the predicted value and the measured value in every component was also very small.

Key words: potato(Solanum tuberosum L.), water, reducing sugar, starch, near infrared spectroscopy

中图分类号: