Internal foreign-matter detection in dark jars
Empty dark jars are imaged from above under dome lighting before filling. Different exposures are merged; deep learning shows foreign matter on the inner floor as a mask and the jar is removed by an air jet.
The anomaly mask is added to the jar record. A jar containing foreign matter is removed through the air jet under PLC control; normal surface variation is accepted.
- 01 Infeed
- 02 Exposure
- 03 Merge
- 04 Anomaly
- 05 Reject
The empty jar enters the station before filling.
- Anomaly regions
- —
- Masked area
- —
Scenes are illustrative simulations; they are not customer references or performance guarantees.
Need
The dark inner floor and bright mouth rim lose detail in a single exposure. Reflections can hide small residue; normal mould traces need to be distinguished from foreign matter.
Solution approach
A cutaway dome spreads reflections; the top camera forms an HDR image from different exposures. Deep learning separates anomaly regions relative to normal inner-surface examples; the mask is overlaid on the jar floor and operator image.
Process steps
- 01
Infeed
The empty jar enters the station before filling.
- 02
Exposure
Different exposures are captured under the dome.
- 03
Merge
Inner-floor detail is merged in the HDR image.
- 04
Anomaly
Deep learning masks the foreign matter.
- 05
Reject
The masked jar is removed 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.