Problem
-
Failure to detect edge cracks, oil stains, insufficient feed, and sunken oxidation on sheet metal strips
-
Failure to detect foreign objects that have fallen into products
Solution
-
Data collection and analysis with Dataguess Inspector
-
Labeling the issues that will be captured in the resulting dataset
-
Enrichment of data
-
Artificial intelligence model improvements, tests and integrations
-
Deployment of the trained artificial intelligence model and transition to a live environment
Result
-
Notifying the operator of cases caught as NG (Not Good) via siren, buzzer, or monitor
-
Performing signaling, automation and integration operations for sending the products detected as NG (Not Good) to the reject line
-
Stopping the production line at the moment of incident capture by sending signals to PLCs