摘要
将二维相关谱与化学计量学结合实现掺杂牛奶与纯牛奶的判别分析。配制40个掺杂尿素牛奶、40个掺杂三聚氰胺牛奶和80个纯牛奶样品,在室温下采集各样品的一维红外光谱。以掺杂物浓度为外扰,分别构建了900-1700 cm^-1和900-1200 cm^-1vs.1200-1700 cm^-1同步红外二维相关谱。在此基础上,分别建立了基于两个区间相关谱掺杂牛奶的多维偏最小二乘判别模型,并与传统一维的偏最小二乘判别模型的分析结果进行对比分析。结果表明：基于900-1200 cm^-1vs.1200-1700 cm^-1二维相关谱模型优于900-1700 cm^-1二维相关谱模型,相关谱的多维偏最小二乘判别模型优于传统一维谱的偏最小二乘判别模型。
Because milk is a complex biological system,it is impossible to extract feature information of adulterants in milk using the conventional one-dimensional（1D） spectroscopy. Comparing with conventional 1D spectroscopy, 2D correlation spectroscopy has higher resolution and better ability in spectrogram analysis. Discrimination analysis of adulterated milk and pure milk is proposed by combining 2D correlation spectroscopy with chemometrics. 40 urea-tainted milk,40 melamine-tainted milk and 80 pure milk samples were prepared, respectively. The mid-infrared spectra of all samples were measured at room temperature. And then the synchronous 2D correlation spectra（900-1700 cm^-1and 900-1200 cm^-1vs. 1200-1700 cm^-1） were calculated to build multi-way partial least squares discriminant analysis（NPLS-DA） models to classify adulterated milk and pure milk. Also, the PLS-DA model was built based on 1D spectra. Comparison results showed that the discriminant models could provide better results using 2D correlation spectra（ 900- 1200 cm^- 1vs.1200- 1700 cm^-1） than 2D correlation spectra（900-1700 cm^-1） and 1D IR spectra.
出处
《中国食品学报》
EI
CAS
CSCD
北大核心
2015年第4期196-200,共5页
Journal of Chinese Institute Of Food Science and Technology
基金
国家自然科学基金项目(31201359
81471698)
天津市自然科学基金项目(13JCYBJC25700
14JCYBJC30400)
关键词
二维红外相关谱
掺杂牛奶
模式识别
判别模型
two-dimensional（2D） infrared correlation spectroscopy
adulterated milk
pattern recognition
discrimi-nant model
