Maturity monitoring of intact fruit and arils of pomegranate cv. ‘Mollar de Elche’ using machine vision and chemometrics

Abstract

Pomegranate fruit cv. ‘Mollar de Elche’ were collected at seven different harvest times. Colour and hyperspectral images of the intact fruit and arils were acquired at each harvest. Physicochemical properties such as total soluble solids, titratable acidity, maturity index, BrimA, internal colour, total phenolic compounds content and antioxidant activity were measured in the juice of each fruit. Relationships between colour (L*, a*, b*) and spectral (720–1050 nm) data obtained from the images of the intact fruit and arils were investigated physicochemical properties using partial least square regression models. Discrimination of the different maturity stages also was carried out using partial least square discriminant analysis models. Similar results were obtained in the prediction of the physicochemical properties using the colour and hyperspectral images of the intact fruit. However, the predictions achieved for the information about the arils were higher using hyperspectral imaging. In the discrimination of maturity stage, the highest accuracies were obtained using hyperspectral imaging, where 95% of intact fruit and 100% of arils where correctly classified. These results indicate the great potential of machine vision techniques, especially hyperspectral imaging, for monitoring the quality of intact ‘Mollar de Elche’ pomegranate fruit and arils.