Neural bruise prediction models for fruit handling and machinery evaluation

Abstract Neural bruise prediction models based on the degree of fruit damage of the most traded fruit species and varieties were developed for prediction of the fruits to be accepted or rejected. The prediction relied on European Community standards. Different models for both quasi-static (compression) and dynamic (impact) loads covering the full commercial ripening period of fruits were developed. A simulation process was developed for gathering the information on laboratory bruise models and load sensor calibrations for different electronic devices (IS-100 and DEA-1, for impact and compression loads, respectively). An evaluation method was also designed for acquiring and gathering the information on the mechanical properties of fruits and the loading records of the electronic devices. The evaluation system allowed for determination of the current stage of fruit handling processes and machinery.