Content area
Abstract
In this study, the potential of near-infrared spectroscopy (NIRS) and hyperspectral imaging (HSI) techniques as non-destructive methods to determine the internal quality attributes of mangoes during ripening were evaluated and compared. A total of 188 mango fruits (cv. Nam Dokmai subcv. Si Thong) were determined for firmness, total soluble solids (TSS), titratable acidity (TA), pH, β-carotene content, and ripening index (RPI) for 8 days. The effect of the position (shoulder, cheek, and tip) of fruit on the changes of fruit quality during ripening and the accuracy of prediction models were also investigated. Fruit spectra were acquired by NIR spectrometer at full wavelength NIR region (800–2500 nm) and HSI system at vis/NIR region (450–998 nm). Partial least square regression was used to develop calibration models using original and pre-treated spectra from both devices. In addition, multiple linear regression (MLR) models were built from specific wavelengths obtained from multifactorial analysis. Non-significant differences of all fruit quality attributes were found between positions at the initial ripening stage while the slightly lower TA and β-carotene content were observed at the shoulder position compared to other sections at the end of the ripening stage. However, the fruit position showed only a slight influence on the prediction performance of the calibration models. NIRS calibration models provided only slightly better prediction performances than HSI calibration models. According to the results, both NIRS and HSI showed potential for quality control in mango sorting.