Research on Apple Internal Quality Classification Based on Near-Infrared Spectroscopy

Abstract

In this study, Luochuan Red Fuji apple in Shaanxi Province was taken as the experimental object, and the sugar content and spectral data were measured and averaged at three locations at the upper distance of the equator. SPXY method was used to divide the data set, MAS, SNV, MC three data preprocessing methods were used, and CARS were used to select the characteristic wavelength, and three classification models SVC, DT and KNN were established. The results show that SPXY + MC + CARS + DT model has the best classification results, and the accuracy rate, accuracy rate, recall rate and F1 score reach 0.955 respectively. In summary, the use of near-infrared spectroscopy technology can be used in Apple's internal quality classification, which improves the basis and reference for the application of Apple's non-destructive testing technology.