Olive Ripening Phase Estimation based on Neural Networks

Color of fruits is a relevant parameter to determine ripeness and optimal harvest time. For olives 6 ripening phases based on skin color distribution have been defined. A widely used method by the olive oil and table olives producers is to inspect the olive surface, and estimate the color and ripening phase visually. This method is simple but it is highly subjective and imprecise. This paper proposes a computational method to estimate the color and ripeness of an olive using digital images. A color scale for olives by means of samples of all ripening phases was developed. To represent the olive color, the histogram of the skin color was proposed as a descriptor. To decide the ripening phase, a classifier based on Neural Networks was implemented. The method allows estimating simply and accurately the olive ripening state, which enables to implement it in real production systems.