GermanyManufacturing2023 – ongoingBosch
AI and generative-AI synthetic images for plant quality inspection
Bosch said AI cut cycle times 15% at its Hildesheim plant and began a generative-AI pilot in two German plants in late 2023
Problem
Training machine-vision systems needs many images of defects, which are rare, and setting up each AI application took six to twelve months.
Approach
Bosch used generative AI to create synthetic images for optical inspection and to tune existing AI models, piloting it in two German plants. Bosch said nearly half of its roughly 230 plants already used AI, and planned to offer synthetic data generation at all sites after the pilot.
Results
Takeaway
Bosch used synthetic defect images to address a shortage of training data. Its reported results came from earlier AI work; shorter set-up times and six- to seven-figure savings per plant remained expectations.
Categories
- Size
- Large enterprise
- Industry
- Manufacturing
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Generative AIComputer vision
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
Bosch reported a 15% cut in cycle times during ramp-up at Hildesheim and a reduction in test time from three and a half to three minutes at Feuerbach. Anomaly detection runs in about 50 plants and machine vision in over 20. Bosch expected generative AI to shorten set-up times and deliver six-figure productivity gains.
Bosch piloted generative AI to make synthetic images for optical inspection in two unnamed German plants. Earlier AI work had cut cycle times by 15% and test time from 3.5 to 3 minutes. Bosch estimated six- to seven-figure yearly savings per plant from generative AI.
Full details are in the original sources.
