The customer's challenge is to reliably detect transparent spoons on reflective metal surfaces to improve quality control
The powder produced at a special nutrition manufacturer is filled into aluminum cans, and then a measuring spoon is placed on the sealed surface. The quality control process involves checking each can to ensure that a spoon is present. When a transparent spoon was introduced to improve the recycling rate, even special sensor-based image processing was unable to reliable identify these spoons on the corrugated, embossed, and reflective metal surface.
To meet the customer's challenge, a 2D snapshot camera is selected that learns using the integrated deep learning software dStudio. Using this web service, neural networks can be trained, in this case, with images of the enclosed spoon in various positions. The camera quickly and reliably checks every product to detect the presence or absence of a spoon. If the camera identifies a “no spoon” image, the system stops automatically. Once the missing spoon has been added, the system continues running without a manual restart.

Benefits:
The manufacturer saves time and money with reliable, accurate, automated quality control. Downtimes are reduced, and quality inefficiencies due to deficient packaging are no longer an issue. The camera can easily be adapted to new products by re-training the neural network by the customer themself. The system is user-friendly and does not require any specialist AI know-how or image-processing knowledge.










