Near-Infrared Spectroscopy for Non-Destructive Prediction of Maturity and Eating Quality of ‘Carabao’ Mango (Mangifera indica L.) Fruit

Abstract

Near-infrared (NIR) spectroscopy was assessed in predicting maturity and eating quality of ‘Carabao’ mango fruit. A total of 1,200 fruits were harvested at the green stage at four different harvest dates [100, 110, 120 and 125 days after flower induction (DAFI)]. Fruits were scanned at the green and table-ripe stage (TRS) using NIR reflectance spectroscopy. The fruits were then measured destructively for the determination of dry matter (DM) content at the green stage, total soluble solids (TSS) and sensory attributes at the TRS. The best calibration models were achieved using partial least square regression (PLSR) analysis for predicting DM, TSS and maturity in the short wavelength region of 700 - 990 nm at 2-nm increment. Principal component analysis-linear discriminant analysis (PCA-LDA) was also used in classifying fruits according to maturity (in terms of DAFI) and eating quality (in terms of overall acceptability or OA). Based on R² values, PLSR models are suitable for quality assurance according to maturity (R² = 0.946, RMSECV = 2.229) and could be used for screening green fruits according to DM (R² = 0.774, RMSECV = 1.091%). The calibration model for predicting TSS (R² = 0.839, RMSECV = 1.282) of ripe fruit using NIR spectra at TRS could be used in research but with caution. For classifying fruits according to DAFI and OA, PCA-LDA gave good results using NIR spectra at the green stage with a success rate of 88% and 86%, 72% and 70% for calibration and validation, respectively.  The findings indicate the potential of near- infrared (NIR) spectroscopy for non-destructive prediction of maturity and quality parameters of mango. The results of the study could serve as the basis for quality control and automatic sorting system for various commodities.