摘要
探讨了利用近红外光谱技术快速、准确地进行生鲜乳中是否含有还原乳鉴别的可行性,并对还原乳检测过程中多元散射校正(MSC)的适用性进行了讨论。文章利用SIMCA判别分析方法,建立了掺入还原乳之生鲜乳的定性判别模型,结果表明:当掺入比例为10%时,正确率达到96.7%,当达到20%以上时,该方法的正确判别率可以达到100%;另外,利用偏最小二乘法(PLS)还建立了还原乳掺入量的定量检测模型(r=0.971,RMSEP=7.76%,RPD=5.13),双尾T检验表明,定量模型预测值与样品真实值间无显著性差异。上述2种方法充分说明了近红外技术可以实现生鲜乳中是否含有还原乳的快速鉴别。
Feasibility of reconstituted-milk identification in raw milk was investigated using near infrared spectroscopy. And the applicability of MSC for reconstituted-milk identification was discussed. The discrimination analysis calibration was developed by SIMCA method, and the result indicated that the accuracy of detection is 100%, when the content of reconstructed Milk is above 20%, while for the 10% reconstituted milk, the accuracy of detection is 96.7%; On the other hand, the quantity models of reconstituted milk were calibrated by partial least squares regression (r= 0. 971, RMSECV= 7.76%, RPD= 5.13), and there were no significant differences between actual value and reconstituted milk prediction value by t test (p=0. 01). All of these suggested that NIRS has good potential to detect adulteration of raw milk with reconstituted milk.
出处
《光谱学与光谱分析》
SCIE
EI
CAS
CSCD
北大核心
2007年第3期465-468,共4页
Spectroscopy and Spectral Analysis
基金
农业部"948"项目(2003-Z74)资助