基于偏最小二乘法建立白酒酒醅近红外分析模型

Based on Partial Least Squares Method to Establish Near Infrared Analysis Models of Fermented Grains

利用近红外光谱分析仪采集酒醅样品的近红外漫反射光谱,基于偏最小二乘法建立检测酒醅中水分、酸度值的近红外光谱分析模型。经内部交叉验证法对模型进行检验,结果为:酒醅中水分的相关系数均大于0.94,酸度的相关系数均大于0.97,证明所建模型测得的酒醅中水分和酸度值与真实值拟合良好。经盲样检测,结果为:模型所测酒醅水分和酸度值均与实际值相匹配且模型稳定性良好,证明所建模型可以实现对酒醅中水分和酸度值的快速检测。

Near infrared diffuse reflectance of fermented grains samples were collected with near infrared spectroscopy, Based on Partial Least Squares Method to Establish Near Infrared Spectral Analysis Model for Measuring moisture and Acidity in Fermented grains.Test the model by the cross validation method,result ：In fermented grains,the correlation coefficient of moisture were greater than 0.94,the correlation coefficient of acidity were greater than 0.97. It proved that the moisture and acidity values of thefermented grains samples measured by the model are well fitted with the true values.Detected by the blind sample, the result ：the moisture and acidity values measuredby the model were matched to the actual values, and model stability is good.It proved that the model could realize the rapid detection of the moisture and acidity in fermented grains.