Together with three other project partners and under the leadership of Fraunhofer IPM, BIA is taking part in the AI innovation competition.
In the surface inspection of components in series production, reliable inspection of components is crucial for quality assurance. Camera-based inspection systems based on AI models offer an efficient solution, but the detection of defects with few sample parts poses a challenge. The project aims to overcome this challenge by developing a generative AI model that synthetically generates image data of defective parts. Based on a small number of real images, the model learns to recognize defects and generate new, plausible images. This expands the inspection system's training data set and improves detection accuracy. Flexible optical inspection systems are used to make the approach applicable across all sectors. In particular, a tunnel inspection system is being developed for two use cases. The approach will be validated in different industries to ensure a successful transfer after project completion.
