A machine developed by RMA-Techniek to classify potatoes using “Deep Learning”.
The potatoes automatically pass under a camera by means of a driven roller conveyor. After that, the screen shows what percentage of which defects are present in the potato sample.
A photo is taken from each side of every potato that passes the camera, due to the rotating transport movement. These pictures are analyzed by a neural network. The network will be able to classify the defects by searching for the differences at the pixel level. The potatoes are classified on external characteristics such as: greenery, growth cracks, scab, infestation, etc. These defects are determined and taught-in per customer.
- Dimensions: 2850x1350x2000mm
- Width conveyor belt: 450mm
- 9 inch control panel
- Conveyor speed adjustable
- Adjustable roller rotation speed
- Completely made of stainless steel 304
- Can be built into an automatic line
- All defects/appearance features can be taught
- Available with dosing system
- This technique can also be used for other products such as: fish, fruit, meat and non-food
If you would like to know more about the possibilities? Feel free to contact us.