Differentiation of deciduous-calyx and persistent-calyx pears using hyperspectral reflectance imaging and multivariate analysis

Deciduous-calyx pears of Korla fragrant pear (Pyrus sinkiangensis Yu) have a significant economic value in Xinjiang Uygur Autonomous Region, China. This study developed a non-destructive method based on hyperspectral imaging using a combination of existing analytical techniques to differentiate the deciduous-calyx pear (DCF) and persistent-calyx pear (PCF). The degrees of circularity of DCP and PCP were extracted according to its morphological characteristic; similarly, the reflectance spectra of DCP and PCP were obtained by hyperspectral imaging technology. Successive projections algorithm (SPA) combined with support vector machine (SVM) established a classification model. The DCF and PCF could be differentiated by SPA-SVM model with accuracy of 93.3% and 96.7% respectively. Our findings suggest that hyperspectral imaging can be applied to non-destructively differentiate pears, and meet the packaging standards.