基于近红外光谱的大鲵黏液粉掺伪的定性定量研究

Qualitative and quantitative study on adulteration of Chinese giant salamander mucus powder based on near infrared spectroscopy

采用近红外光谱扫描不同掺伪样品,用偏最小二乘线性判别法分别建立掺伪大鲵黏液粉的2分类(纯大鲵黏液粉和掺伪样品)和5分类(纯大鲵黏液粉、掺入土豆淀粉、大鲵肉粉、大鲵蛋白粉和大鲵皮粉)定性判别模型;用偏最小二乘回归分别建立4类掺伪大鲵黏液粉的定量校正模型。结果表明,原始光谱经一阶导数+标准变量变换预处理后,所建2分类和5分类小二乘线性判别定性模型的校正集和预测集均无误判数;大鲵黏液粉掺入土豆淀粉、大鲵肉粉、大鲵蛋白粉和大鲵皮粉的定量预测模型能够较好地验证黏液掺伪成分。

In this paper near infrared spectroscopy was used to scan different adulterated samples,and partial least squares linear discriminant method was used to establish qualitative discriminant models for two classifications of adulterated giant salamander mucus powder(pure giant salamander mucus powder and adulterated sample)and five classifications(pure giant salamander mucus powder,potato starch,giant salamander meat powder,giant salamander protein powder and giant salamander skin powder)respectively.Regression models were established for 4 kinds of adulterated giant salamander slime powder.The results showed that the calibration set and prediction set of PLS-DA qualitative model for two and five classifications were correct after the first derivative and standard variable transformation pretreatment.The quantitative prediction model of giant salamander mucus powder mixed with potato starch,giant salamander meat powder,giant salamander protein powder and giant salamander skin powder can correctly verify the mucus adulteration ingredients.