Defect classification and sorting of tablets and capsules
Tablets and capsules travel in single file along a vibrating channel. A colour image from above is inspected for shape, colour and print; defective products are diverted to a reject chute by an air jet.
Product class and defect area are sent to the operator display and line PLC; the reject decision is matched to product travel and sent to the air jet.
- 01 Infeed
- 02 Acquire
- 03 Classify
- 04 Sort
The product enters the field of view in single file.
- Defect area
- —
- Detected products
- —
Scenes are illustrative simulations; they are not customer references or performance guarantees.
Need
Product orientation and reflections vary; a capsule seam can resemble a fracture. Overlapping products disturb the assessment of the shape and print of an individual item.
Solution approach
An overhead colour camera and diffuse lighting reduce surface glare. Deep learning classification evaluates fracture, colour and print patterns within the product region; shape and overlap checks contribute to the air-jet decision.
Process steps
- 01
Infeed
The product enters the field of view in single file.
- 02
Acquire
A colour image is captured from above under diffuse light.
- 03
Classify
Shape and surface patterns are classified.
- 04
Sort
A broken tablet is sent to the reject chute by the air jet.
Related applications
Let us assess your application together
We test your sample in our laboratory and report on resolution, optics and lighting selection.