Abstract :
The consumer at the retail outlet rejects tomatoes harvested before reaching physiological ripeness, as the produce does not ripen properly. Although many packing houses have inserted vision systems for tomato sorting and sizing, it is impossible to distinguish between fruits that will ripen and fruits that will not at the harvest day, as all the fruits are green. A tomato maturity predictive sensor was developed to be used at packing houses for detecting unripe tomatoes that will never turn red. Spectral and colour measurements were taken from green-mature tomatoes just after harvest. Colour was measured daily and used as an index for classifying first-day spectral data into ripe or unripe. The groups were analysed by discriminant analysis in order to obtain the best discriminant wavelengths. Those spectral bands were used by a multi-spectral imager for predicting those tomatoes that will never ripen with an accuracy of over 85%.
