Identification of True and False Aksu Apple Based on NIRS and PLS-DA

Abstract

A method for the identification of true and false Aksu Red Fuji apple varieties based on near-infrared spectroscopy (NIRS) and partial least squares discrimination (PLS-DA) was proposed and established. Authentic Xinjiang Aksu Red Fuji and ordinary Red Fuji (Shandong and Shanxi Red Fuji) were purchased from supermarkets. There were 42 samples in each variety, with totaling 126 samples. Near infrared diffuse reflectance spectra of all samples were collected in the range of 4000-12000 cm−1. On the basis of studying the characteristics of NIR spectra, the principal component analysis (PCA) was carried out. It is pointed out that PCA method cannot fully realize the identifcation of three kinds Red Fuji Apples. Finally, the partial least squares discriminant analysis model of three kinds Red Fuji Apples was established, and the correct discrimination rates of the three kinds Red Fuji Apples were 100% for samples in calibration set and prediction set.