Application of VIS-SW-NIR Spectroscopy for Detecting Adulteration in Fresh Minced Beef
Xiaoyu Tian,Yang Xu,Yankun Peng,Xiuying Tang,Shibang Ma
DOI: https://doi.org/10.13031/aim.20131591010
2013-01-01
Abstract:The purpose of this study is to evaluate the potential of visible and short wave near infrared (VIS-SW-NIR) spectroscopy, for the rapid, nondestructive detection of beef adulteration. Visible and short wave near infrared (VIS/SW-NIR) spectroscopy was acquired from pure minced beef pork and beef liver, beef and pig fat trimming as well as the mixture samples in the form of minced beef adding different proportion of others, respectively. Partial least squares regression (PLSR) models were established to predict concentration of different kinds adulterants, the correlation coefficient R reached 0.9729, 0.9919, 0.9598, and 0.9445, the root mean square prediction error (RMSEP) were 5.9367, 4.4131, 10.0220 and 10.9009, respectively. According to the average spectral curves of samples with different adulterating level, linear discriminant analysis (LDA) model and support vector machine (SVM) models were built to identify whether the foods were adulterated, and furthermore, classify the level of adulterated level, by selecting some important characteristic wavelength most related to the difference of each adulterants in the minced beef, respectively. The classification accuracy from both types of models was quite high, especially the total prediction accuracy of LDA model reached 94.4, 84.7, 93.3 and 72.2 percent, respectively. The current study demonstrated that the VIS-SW-NIR spectroscopy can be used to detect and classify the amount and level of adulterants added to the minced beef with acceptable precision and accuracy.