借助近红外光谱技术,以江西省永丰官山林场楠木种子园的64份闽楠叶片为材料,采用传统的化学分析方法测定闽楠叶片样品的氮元素含量作参考值,同时采集闽楠叶粉末样品的近红外光谱图。运用化学计量学软件NIRCal,选定建模方法、建模波段和预处理方法,建立最优测定模型。另随机抽取10份未知样品对模型进行检验,结合配对样本T检验进行评价。结果表明:用主成分回归方法(PCR)建立的模型效果最好,其校正集相关系数Rc为0.912,校正集均方根误差RMSEC为1.098,交互验证集相关系数Rv为0.897,交互验证集均方根RMSEV为1.192。外部验证结果显示,预测值与实测值的相对偏差范围在0.070~0.705之间,且配对样本T检验结果显示P值为0.116,大于0.05,无显著差异。该方法可用于闽楠营养水平和优质选种的大批量快速检测。

Using near-infrared spectroscopy,64 leaves of P.bournei in the P.bournei Seed Garden of Guanshan Forest Farm in Yongfeng,Jiangxi Province were used as materials,and the nitrogen content of P.bournei leaf samples was determined by the traditional chemical analysis method as a reference value.At the same time,the near-infrared spectrum of the P.bournei leaf powder samples was collected.Using the chemometric software NIRCal,the optimal determination model was established by selecting the modeling method,modeling band,and pre-treatment method.In addition,10 unknown samples were randomly selected to test the model,and the paired sample T test was used to evaluate the model.The results showed that the model established by the Principal Component Regression (PCR) method had the best effect;its related coefficient of calibration (Rc) was 0.912;the Root Mean Square Error (RMSEC) of calibration was 1.098;the related coefficient of validation (Rv) was 0.897;and the Root Mean Square Error of validation(RMSEV) was 1.192.The external verification results showed that the relative deviation between the predicted value and the measured value ranged from 0.070 to 0.705,and the paired sample T-test results showed that the P value was 0.116,greater than 0.05,and there was no significant difference.It can be used for large-scale rapid detection of the nutrient level and high-quality seed selection of P.bournei.