A team of scientists at the Valencian Institute of Agrarian Research (IVIA) located in Spain have developed a computer vision-based machine that can detect and segregate spoilt oranges. They have also developed two other machines for checking the quality of mandarins and for assisting citrus fruit pickers in the field.
The machine classifies citrus fruits on the production line. Credit: IVIA
Traditionally oranges that are spoilt are picked out by passing the oranges under ultraviolet light in a dark room. The essential oils that are present in the rind of the spoilt oranges are lit up using florescence. Since ultraviolet light is harmful, this is done in stringently timed shifts.
The machine that has been developed by researchers uses computer software to interpret the image that is in front of it and behaves accordingly ensuring segregation of high quality and spoilt oranges. The team has also developed another machine that can categorise citrus fruits based on their quality, colour, and the damage type at a speed of 20 to 15 fruit pieces per second. This kind of classification is important where high quality fruits should be sent for premium markets and second class quality fruits that have only slight damage but are still edible must be sent to other markets.
Another machine, which has been developed inspects the quality of ready to eat mandarin segments at a speed of 28 segments per second. All the segments are passed through a conveyor belt and the broken segments are separated. Magnetic image resonance is also being used to inspect the internal quality of the fruit. IVIA scientists have also developed a self-run tractor that helps in picking citrus fruits by using sensor and computer vision system.