Only top quality is accepted by the AI

Osnabrück. In an innovative project in collaboration with potato manufacturer Pahmeyer, the technical college has researched the automation of the quality assurance process for potato pancakes in food production using artificial intelligence. The buffers from the factory are subject to a strict quality assurance process before they reach the customer. Products with the smallest defects, such as uneven browning or irregular shape, are sorted out by Pahmeyer employees directly during production. A class of electrical engineering students from Osnabrück had the task of automating this process. They used state-of-the-art technologies from the field of artificial intelligence. Using hundreds of images of potato buffers, they trained neural networks to classify new products from production and sort them out if necessary. The students came to the conclusion that the trained models delivered amazingly good results, but still made too many wrong decisions. As a result, the system is not yet sufficiently reliable for use in production. Pahmeyer will therefore continue to need the experience and trained eye of its employees in the near future.

- Stefan Sayk (Director of Studies, Electrical Engineering, Computer Engineering | Berufsbildende Schulen Osnabrück)