不同近红外光谱仪在茶叶品质检测中的应用研究

分别应用NIRSystems 6500和InfraXact Lab两种不同型号近红外光谱仪对茶叶中茶多酚和咖啡碱含量进行定标建模,确定了在改进偏最小二乘法定标技术下的最优光谱数据预处理方式,分别建立了茶多酚和咖啡碱的优化定标模型.NIRSystems 6500和InfraXact Lab两种型号近红外光谱仪所建立茶多酚的最优定标模型目标函数f值分别为96.07%和95.96%,咖啡碱的最优模型目标函数f值分别为95.32%和91.47%,定标效果良好.利用验证集样本对模型进行了检验,茶多酚和咖啡碱的预测相对分析误差(RPD)均大于3.结果表明,两种仪器所得到的模型定标效果均良好,预测精度高,都可用于实际检测.NIRSystems 6500型近红外光谱仪的定标效果和稳定性优于InfraXact Lab型近红外光谱仪.

Application of Different Near-infrared Spectrometers in Tea Quality Inspection

Two different types of NIRSystems 6500 and InfraXact Lab near-infrared spectrometer were used to establish calibration models for the tea polyphenol and caffeine of tea. The optimal spectral data pretreatments were confirmed with the modified partial least squares calibration technology for each spectrometer. The optimal calibration models of tea polyphenol and caffeine in tea were established. The tea polyphenol target value f were 96.07% and 95.96% of two different types near-infrared spectrometer,respectively. The caffeine target value f were 95.32% and 91.47%,respectively. The optimal calibration models were tested by validation sets,the result of RPD was greater than 3. The results showed that the calibration effects of 2 models were good,and there were high prediction accuracy. Both of two types of pretreatments can be used in actual testing,but the calibration effect and stability of the NIRSystems 6500 near-infrared spectrometer was better than the type of InfraXact Lab.