Rapid identification and quantification of intramuscular fat adulteration in lamb meat with VIS–NIR spectroscopy and chemometrics methods

Meat adulteration can be one of the main reasons of human’s healthy and safety problems. Therefore resolving this problem is a significant issue in food industry. VIS–NIR spectroscopy in this work was used as nondestructive technique to classify and evaluate the quantity of intramuscular fat in minced lamb meat. There were totally 110 samples and every sample weighed 10 gr. adulterated samples were prepared manually with 5%, 10%, 15%, and 20% (w/w) adulteration levels. Principle Component Analysis and Linear Discriminant Analysis (LDA) models were applied with different preprocessing methods to separate unadulterated and adulterated samples in to two and five class datasets. The best results of LDA model was 86.2% and 100% accuracy with Savitzky–Golay smoothing preprocessing for five and two class datasets, respectively. Partial Least Squares Regression model was built under cross-validation and external validation testing to quantify the adulteration level of samples. The best outcome of this model was with SNV with Correlation Coefficient of prediction Rp2 = 76.51% and root mean square error of prediction RMSEP = 0.76. Then Ultraviolet–Visible-Near infrared spectroscopy can be used safely as non-destructive technique for detection of adulteration in meat industry.