Overview Deep Learning improves yields in high volume can manufacturing by automatically rejecting bad cans. The Snap Platform alerts in real time when successive defects occur to help operators fix problems immediately.
Metal can manufacturing processes are prone to small defects and anomalies, often on the inside of the can. Defects tend to come quickly and affect all successive cans, so detecting defects as quickly as possible for intervention is important to maintain production quality. A difficult imaging and inspection problem combined with very high throughput rates makes can inspection impossible for humans to do inline.
The Overview Solution
Our Deep Learning platform is able to make lightning quick decisions over multiple types of defects on the inside of the can.
How does the
Deep Learning Work?
Easy integration combined with state of the art models
A flexible imaging setup allows comprehensive imaging of cans
Full inspection of the inner walls of each can, from the top edge to the bottom
Deep learning models look for longitudinal scratches, bent metal, and stamp defects
Defective cans are immediately ejected from the production line
Consistent, reliable, 100% in-line coverage inspection removes slow end of line manual inspectors. Increases overall factory efficiency.
Demo the Snap Platform to see how we improve your factory efficiency