枣醋发酵液中还原糖含量和酒精度的可见–近红外高光谱快速定量分析

Determination of Reducing Sugar and Alcoholic Strength in Jujube Vinegar Fermented Liquid by Hyperspectral Image Technology

本文利用可见–近红外高光谱检测技术对枣醋发酵过程中还原糖含量和酒精度进行了定量分析,并通过偏最小二乘法(PLS)建立定量分析模型,同时采用无信息变量消除法(UVE)、竞争性自适应加权算法(CARS)和遗传算法(GA)对整个谱区进行光谱特征波长变量筛选,以决定系数(R2)、预测标准偏差(RMSEP)、相对分析误差(RPD)以及最佳主因子数作为模型质量的评价指标。其中用CARS法挑选波长后对模型的优化效果最佳,还原糖含量和酒精度CARS-PLS模型的R2分别达到0.9045和0.8993,RMSEP为1.3635和1.2878,RPD为3.24和3.58,最佳主因子数为7和8。结果表明:进行变量筛选可提高枣醋酿造过程中还原糖含量和酒精度模型准确度和稳定性,降低解析难度,达到优化模型的作用,可见–近红外高光谱检测技术可实现枣醋发酵液中还原糖含量和酒精度的快速、定量、精确分析。

Reducing sugar and alcoholic strength content in jujube vinegar fermented processing was carried out by hyperspectral image technology, and quantitative analysis model was established by partial least squares method, while the spectral characteristic wavelength of the spectral region was screened by uninformative variables elimination (UVE), competitive adaptive reckons (CARS) and genetic algorithm (GA). Determination coefficient R2, root-mean-square error of prediction (RMSEP), relative percent deviation (RPD) and the best principle factors were employed as evaluation index-es. The optimization effect of CARS was the best, R2 of optimization mode were 0.9045 and 0.8993, RMSEP were 1.3635 and 1.2878, RPD were 3.24 and 3.58, and best principle factors were 7 and 8. The results showed that: variable selection can improve accuracy and stability of the model of sugar and alcoholic strength content quantitative analysis, reduce analytical difficulty, and optimize the model. It is feasible that reducing sugar and alcoholic content of jujube vinegar fermented liquid can be analyzed rapidly, quantitatively and accurately by hyperspectral image technology.