基于二维相关近红外光谱技术的牛奶掺杂识别方法的研究-生物医学工程专业论文.docxVIP

ABSTRACT The safety of dairy products is a global problem, while the traditional test means of dairy products can’t meet the needs of production quality control and safety control. With a lot of advantages such as efficient, fast, convenient, without pre-treatment, etc. Near-infrared(NIR) spectroscopy makes up the deficiencies of the traditional detection methods and was applied to the field of milk testing. Two-dimensional(2D) correlation spectroscopy has higher resolution than the conventional one-dimensional spectroscopy, so NIR spectroscopy combined with 2D correlation techniques and a variety of pattern recognition methods are applied to discriminant analysis for adulterated milk in this thesis. The range of wave number was fixed to 4200~4800cm-1 through analyzing the one-dimensional NIR spectrum of adulterated milk(urea, melamine). 2D correlation synchronous spectrum was chosen after compared with 2D correlation asynchronous spectrum. 2D correlation synchronous spectrum combined with pattern recognition was used for discriminant analysis of adulterated milk. Mainly two methods, (1)Unfold-PLS, 2D correlation spectrum matrix of each sample was unfolded along the direction of the wave number and combined with ordinary Partial Least Squares(PLS)for discriminant analysis. (2)K-OPLS, 2D correlation spectrum matrix(three-dimensional data) combined with K-OPLS were used for discriminant analysis. The results showed that these two discriminant models could be effective for milk adulterated or not. In order to improve the HYPERLINK /efficiency efficiency of modeling, feature information of the huge amount of the 2D correlation spectrum data was needed. Mainly two methods, (1)Five statistical parameters(mean, standard deviation, center of gravity, skewness, and kurtosis) were extracted from 2D correlation synchronous, then Partial Least Squares Discriminant Analysis(PLS-DA), Back Propagation(BP) neural network and Probabilistic Neural Networks(PNN) were used to establis