基于近红外光谱技术测定稻谷水分含量研究

本文采用近红外光谱技术结合化学计量学方法，建立稻谷水分含量测定的快速分析方法。试验选取江苏省不同地区的两年内197份稻谷样品作为建模集样品，对其进行化学分析和图谱扫描处理，通过近红外化学计量学软件初步建立稻谷水分含量的预测模型。建模结果显示运用PLS（偏最小二乘法）建立的分析模型预测效果最优，决定系数（R2）高达0.9689，交互验证标准差（SECV）为0.3434，选取24个未知样品作为验证集样品，验证决定系数（R2）高达0.9806，预测标准差为0.0933。结果表明，近红外光谱技术可以用于稻谷水分含量的快速测定。

The research on determination of moisture content of rice based on the near infrared spectroscopy

Abstract In this paper, established a rapid analysis method for determination of rice moisture content using near infrared spectroscopy(NIR) technology combined with chemometrics methods. 197 rice samples were collected as calibration samples in different parts of the two years in Jiangsu province, with the process of chemical analysis and map scanning, established the forecast model of rice moisture content through near infrared chemometrics software.Modeling results indicate that model established by using the PLS (partial least square) had the optimal predictive effect, the determination coefficient (R2) was 0.9689, standard error of cross-validation(SECV) was 0.3434, 24 samples was selected as validation set, the determination coefficient (R2) was 0.9806, standard error of forecast was 0.0933.The results showed that NIR can be used in the rapid determination of rice moisture content.