基于近红外光谱分析技术的大米水分含量快速检测研究

Rapid Detection of Rice Water Content Based on Near Infrared Spectrometry

为实现大米水分含量的快速、无损检测,提高检测效率,收集来自5个不同地区共120份大米样品,基于近红外漫反射光谱分析技术建立大米水分含量的定量模型。结果表明:多元散射校正(MSC)预处理方法结合主成分回归(PCR)分析建立的定量模型性能最优,其校正集相关系数及校正集均方根标准差分别为0.9488和0.0048,预测集相关系数及预测集均方根标准差分别为0.9362和0.0065。建立的近红外光谱定量模型不仅具有快速、无损的特点,而且预测结果具有较高的可靠性和准确性,可用于大米水分含量的快速检测。

In order to achieve rapid and non-destructive detection of rice moisture content and improve testing efficiency.A total of 120 rice samples from 5 different regions were collected,and a quantitative model of rice moisture content was established based on near-infrared diffuse reflectance spectroscopy.The results showed that the quantitative model established by the multivariate scatter correction(MSC)preprocessing method combined with the principal component regression(PCR)algorithm gets the best performance.The correlation coefficient of the calibration set and the standard deviation of the root mean square of the calibration set were 0.9488 and 0.0048,respectively,and the correlation coefficient of the prediction set and the standard deviation of the root mean square of the prediction set were 0.9362 and 0.0065,respectively.The near-infrared spectrometry quantitative model established is not only fast and non-destructive,but also had high reliability and accuracy in the prediction results,which could be used for rapid detection of rice moisture content.