A comparison between NIR and ATR-FTIR spectroscopy for varietal differentiation of Spanish intact almonds

The rapid and easy classification of almond varieties with similar morphology, different quality properties and, in most cases, different prices is interesting to protect both the almond industry and the consumers from fraud. Therefore, in this work, intact almond kernels from four Spanish varieties ('Guara', 'Rumbeta', 'Marcona' and 'Planeta') were analysed using both near infrared (NIR) and attenuated total reflectance Fourier-transform infrared (ATR-FTIR) spectroscopy. After spectra measurement, NIR and ATR-FTIR spectral data were pre-treated and employed to construct two classification methods (partial least square-discriminant analysis (PLS-DA) and quadratic discriminant analysis (QDA)) in order to check their ability to classify almonds according to their variety. The best overall accuracies (94.45%) were obtained with the PLS-DA model of ATR-FTIR and the QDA model of NIR data. These results confirm that both spectroscopic techniques, if the optimal statistical model is selected, are powerful tools to reliably discriminate almonds according to their varieties.