采用近红外漫反射光谱法分析牛奶成分

为提高牛奶成分化学分析的效率，研究了近红外漫反射光谱在牛奶主要成分分析中的应用；讨论了光在牛奶中的传播方式和相关测量方法的选择；采用偏最小二乘(PLS)建立校正模型，比较了各种预处理方法和不同谱区的建模效果，其中脂肪的相关系数为0．99，标准偏差为0．048 g/dL，蛋白质的相关系数为0．92，标准偏差为0．102g/dL结果表明，近红外漫反射光谱测量可以满足牛奶主要成分的测量要求。

Composition Analysis of Milk by Using Near Infrared Diffuse Reflection Spectroscopy

In order to improve the analysis efficiency, this paper studies the method of analyzing the composition of milk by using nearinfrared diffuse reflection spectroscopy and discusses the propagation of the light in the milk and some relevant measuring methods. The model is built by partial least-squares (PLS),pretreatment methods and different spectral regions are compared. The best correlation values of fat and protein are 0.99 and 0.92 respectively, the best root mean square error of prediction (RMSEP) values of fat and protein are 0.048 g/dL and 0.102 g/dL respectively. The study shows that near infrared diffuse reflection spectroscopy is a potentially useful technique for evaluating the composition of milk.