Defect and class sorting of multilayer ceramic capacitors
Capacitors are placed on a glass rotary table and imaged under an overhead camera. Ceramic surfaces and metal electrode ends are classified with deep learning; a non-conforming part is separated by an air jet.
Part class is recorded with the table position. When the non-conforming capacitor reaches the jet point it is blown into the reject chute; class distribution is shown on the display.
- 01 Index
- 02 Acquire
- 03 Classify
- 04 Separate
The glass table brings the capacitor under the camera.
- Class code
- —
- Defect regions
- —
Scenes are illustrative simulations; they are not customer references or performance guarantees.
Need
A fine crack on the ceramic body and a reflection on a metal end can form similar lines. Orientation and electrode edges change the appearance of the small part in the image.
Solution approach
The capacitor on the glass table is imaged from above with micro optics. Oblique lighting reveals ceramic cracks and electrode boundaries; deep-learning classification separates sound parts, cracks and electrode defects. Table position is linked to the reject jet.
Process steps
- 01
Index
The glass table brings the capacitor under the camera.
- 02
Acquire
Oblique light reveals the body and electrodes.
- 03
Classify
The image is compared with sound and defect classes.
- 04
Separate
The defective capacitor is sent into the chute by an 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.