山茶油多元掺假近红外模型的建立与研究

Establishment and study of camellia oil multiple adulteration near-infrared model

中文关键词:  山茶油  掺假检测  近红外光谱  多元掺假  模型优化

英文关键词:camellia oil  near infrared spectroscopy  adulteration detection  multiple adulteration  model optimization

中文摘要:

      目的 建立基于近红外光谱结合偏最小二乘法的山茶油、花生油和玉米油多元掺假模型。方法 通过对比不同建模光谱波段、预处理方法对模型进行优化研究, 确定最优的山茶油、花生油和玉米油多元掺假模型。结果 所建模型性能指数均大于0.99, 校正集和预测集的均方差在0.6以内。经未知掺假样品的外部验证, 预测值与实测值之间有较好的相关性, 二元掺假模型预测值的相对误差在1%以内, 三元掺假在6%以内, 验证结果良好。结论 近红外光谱结合偏最小二乘法的检测技术快速、有效、环保, 可用于定量检测山茶油的掺假。

英文摘要:

      Objective To establish a multiple adulteration model of camellia oil, peanut oil and corn oil based on near-infrared spectroscopy combined with partial least squares method. Methods The model was optimized by comparing different spectral bands and pretreatment methods. The optimal multiple adulteration model of camellia oil, peanut oil and corn oil was determined. Results The performance indexes of the models were all greater than 0.99, and the mean variance of the correction set and prediction set was within 0.6. After the external verification of unknown adulteration samples, there was a good correlation between the predicted value and the measured value. The relative error of the predicted value of the binary adulteration model was within 1%, the ternary adulteration was within 6%, and the verification results are good. Conclusion Near-infrared spectroscopy combined with partial least square method is fast, effective and environmentally friendly, which can be used for quantitative detection of adulteration of camellia oil.