| 摘要: |
| 研究使用DA2700型近红外光谱仪采集了112个湿加松Pinus elliottii×P.oaribaea松针粉末样本的光谱数据。结合实际测定值，采用偏最小二乘（PLS）回归法并选择最佳光谱预处理方法和最佳主成分数，建立湿加松松针黄芪苷含量的近红外快速预测模型。结果表明：当采用一阶导数（FD）+标准正态变量转换法（SNV）对光谱数据进行预处理，主成分数为6，此时模型的预测效果最好，校正集相关系数（RC）和交互验证集相关系数（RV）分别为0.8082和0.7109。校正集均方根误差（RMSEC）和交互验证集均方根误差（RMSEV）分别为1.9314和2.3988，说明模型的预测效果较好。利用外部验证集对模型进行验证，得到模型的外部验证相关系数R=0.8129，预测均方根误差RMSEP=2.9738。 |
| 关键词: 湿加松 黄芪苷 近红外 预测模型 |
| DOI： |
| 投稿时间：2021-05-27修订日期：2021-08-05 |
| 基金项目:台山市红岭国家湿地松、杂交松良种基地 2020 年中央财政林林木良种繁育补助项目 |
|
| 基于近红外光谱技术建立湿加松针叶黄芪苷含量的预测模型 |
| peng guanming,lv xinxin,mao jipeng,ou huiling,xie nuo,li fuming |
| (Taishan Hongling Seed Orchard;South China Agricultural University) |
| Abstract: |
| In this study, the DA2700 near infrared spectrometer was used to collect the spectral data of 112
wet pine needle powder samples. Combined with the actual measured value, the near infrared rapid prediction model of astragalin content in wet pine needles was established by partial least squares (PLS) regression method and selecting the best spectral pretreatment method and the best principal component fraction. The results show that the prediction effect of the model is the best using the combination of first derivative (FD) and standard normal variable transformation (SNV) method to preprocess the spectral data, and when the principal component fraction is 6, the correlation coeffificient of correction set (RC) and cross validation set (RV) was 0.808 2 and 0.710 9 respectively. The RMSEC and RMSEV was 1.931 4 and 2.398 8, respectively, indicating that the prediction effect of the model was better. The external validation set was used to verify the model, and the correlation coeffificient of external validation was R=0.812 9 with the root mean square error of prediction was RMSEP = 2.973 8. |
| Key words: Prediction Model of Astragalin Content in Pinus elliottii × P. caribaea Needles Based on Near Infrared Spectroscopy |
