Olive sorting line
Requirement
Manual sorting of defective olives was slow and produced results that varied between operators. The facility wanted to increase capacity during the harvest season while maintaining a consistent quality standard.
Our approach
A ZEYBEK-20V1-2500 optical sorting machine was manufactured. Each olive is classified using an LED-lit imaging tunnel, two industrial colour cameras and a deep-learning model running on a Linux-based processor; a pneumatic ejector group with 20 independent channels directs defective olives to the reject outlet. The model was trained for the olive varieties processed by the facility and delivered with the capability to train new profiles on site.
Outcome
It operates at a minimum capacity of 2,500 kg/h with a detection rate of at least 95% for the defect classes verified in the factory acceptance test.
Technologies used
- ZEYBEK-20V1-2500
- NVIDIA Jetson
- Beckhoff PLC
- Festo
- Basler