基于2T2D相关特征切谱掺杂牛奶判别

Discrimination of Adulterated Milk Based on Two-Trace Two-Dimensional Correlation Feature Slice Spectra

提出并建立一种基于2T2D相关特征切谱掺杂牛奶判别方法。制备两个品牌纯牛奶和掺杂三聚氰胺牛奶(0.02~2 mg·mL^(-1))各80个样品,共160个,以浓度为外扰,构建同步2T2D相关红外谱,并在纯牛奶和三聚氰胺特征峰位置对同步谱进行切谱处理,得到每一个样品的特征切谱。分别基于2T2D相关特征切谱、同步谱和室温下一维中红外光谱建立两个品牌融合的掺杂牛奶判别模型。结果表明:所提出的2T2D相关特征切谱判别方法不仅压缩了数据,而且有效提取了特征信息,提高了判别准确率。

A discriminant method of adulterated milk was proposed and established using two-trace two-correlation(2T2D)correlation feature slice spectra.A total of 160samples were prepared including two brands of 80pure milk and melamine-tainted milk(0.02~2mg·mL^(-1)).Synchronous 2T2Dinfrared correlation spectrum of each sample was calculated under the external perturbation of concentration.The characteristic slice spectra of each sample were extracted from synchronous 2T2Dcorrelation spectrum at characteristic peaks of milk and adulterants.the discrimination models of two brands of adulterated milk were established using 2T2Dslice spectra,2T2Dcorrelation spectra and one-dimensional infrared spectra at room temperature,respectively.The results showed that the proposed method not only compressed the data,but also effectively extracted the characteristic information of the components to be analyzed,and improved the accuracy of discrimination.