An Automated System for Sorting of Freshly Harvested Tomato Fruits

Abstract

Fruit sorting determines market value. Farmers and traders commonly use physical-eye inspection and handpicking for sorting, but this is labour-intensive and ineffective. This research work aims to develop a sensor-based automated system for sorting freshly harvested tomato fruits. The automated system sorts tomato fruits into small, medium, and big sizes for market value. To evaluate the system performance, 115 fruits were machine-sorted and compared to eye-inspection and physical measurement. Physical measurement was done by measuring the minor, intermediate, and major diameters of each fruit with a Vernier calliper. While the eye-inspection was carried out by manual human examination with the eye. Results show average percentage error between physical measurement and automated sorting is 10.264%, which implies 89.736% accuracy. The influence of conveyor speed at three levels (2.8, 3.4, and 3.9) cm/sec on overall system performance was evaluated, and the optimum speed of 3.4cm/sec was obtained.