近红外光谱法快速检测金银花中水分的含量

Rapid Determination of Moisture in Honeysuckle by Near-infrared Spectroscopy

摘要: 目的应用近红外光谱法和数据分析软件，对金银花中水分含量进行快速测定。方法利用甲苯法测定样品中水分的含量，运用偏最小二乘(PLS)法建立其含量与NIR光谱之间的多元校正模型，对未知样品进行含量预测。结果建立的水分校正模型相关系数(R2)、估计误差均方根(RMSEE)、相对分析误差(RPD)分别为0.933， 0.18%，3.86。经外部验证，校正模型的预测均方差(RMSEP)、平均回收率分别为0.216， 98.9%。结论此方法具有快速简便、准确无损的特点，可应用于金银花中水分含量的快速检测。

Abstract: OBJECTIVE To determine moisture in honeysuckle rapidly by near-infrared spectroscopy and data analysis software. METHODS Toluene method was used as a reference method to determine the content of moisture in honeysuckle. Multivariate calibration model based on PLS algorithm was developed to correlate the spectra and the corresponding values determined by the reference method. RESULTS The correlation coefficients (R2), the root-mean-square error of estimated(RMSEE) and the RPD of the calibration model for moisture were 0.933, 0.18% and 3.86, respectively; the root-mean-square error of prediction(RMSEP) and the average rate of recovery were 0.216 and 98.9%. CONCLUSION The method is fast and convenient. The correction model could be used to predict moisture in honeysuckle rapidly.