Physicochemical composition of persimmons: A non-destructive determination approach

Résumé

Fruit from eight persimmon (Diospyros kaki) cultivars were evaluated for physicochemical quality at physiological maturity, and classification models were developed using near-infrared spectroscopy (Vis-NIR). Physical traits (mass, diameter, height), color (L*, a*, b*, chromaticity), firmness, soluble solids content, pH, titratable acidity, moisture, and dry matter were evaluated. Vis-NIR Spectra (310–1100 nm) were collected on a subset of five cultivars (n = 20 fruits), preprocessed with SNV, MSC, and Savitzky–Golay derivatives, and analyzed by PCA and PLS-DA; ANOVA supported comparisons. Pomelo showed higher fruit mass and diameter. Giombo and Taubaté had higher L* values and SSC and TA did not differ significantly between cultivars. Moisture content was highest and DM lowest in hybrid 82, while Giombo showed the lowest firmness. PCA explained 91–97% of spectral variance without clear separation across cultivars. PLS-DA achieved the best validation for Rama Forte (R²_V = 0.75; RMSE_V = 0.21) and Giombo (R²_V = 0.68; RMSE_V = 0.25). Vis-NIR spectroscopy successfully classified two persimmon cultivars, which can possibly be used for authentication of the product offered to consumers.