The emergence of new technologies focusing on “image analysis” contributes significantly to the assessment of fruit quality based on objective and non-destructive features. In this investigation, the ‘Mejhoul’ date fruit cultivar was subjected to freezing at -10°C and -18°C and stored for 6 months. Its quality was evaluated according to texture features extracted from images acquired using a digital camera and flatbed scanner. The extraction process was carried out according to an internal procedure using MaZda software. Then, the extracted features were used as inputs for pre-established algorithm groups within WEKA software to classify frozen date fruit after 0, 2, 4, and 6 months of storage. Accordingly, reducing these features using the “Best-First” method allowed for a selective ranking of about twenty accurate features that were submitted to four classifier groups of algorithms: Bayes, Functions, Lazy, and Trees.

The results allowed for the extraction of a hundred texture features that differ depending on “storage temperature” and “storage period”. Furthermore, high accuracy levels for the classification of ‘Mejhoul’ date fruit were obtained for each storage period based on the selected features, and slight differences were noted between the algorithms used. In future, physicochemical attributes will be added to the developed models to correlate with image features and predict the behaviour of date fruit under storage.