Error detection
in the paintwork
An automotive supplier checks the results of a painting robot manually, at great expense in terms of time and with a high susceptibility to errors.
Procedure
- Realisation of a two-day workshop at the customer's premises
- Analysis of known error patterns
- Development of relevant variables on the basis of log data
- Development of a learning image recognition system using Amazon SageMaker
- Creation of classification models to recognise the error patterns
- Training of machine learning algorithms (Google Tensorflow) using historical data.
Result
- Recognition of 5 known error patterns with 96% Accuracy
- Detection of 2 unknown error patterns with 87% Accuracy
- Use of the Error detection components planned for the next product line
Find out now with just a few questions whether your problem can also be solved by AI.SEE™.
AI.SEE™ uses artificial intelligence to enable manufacturers to carry out non-destructive, reliable and automated inspections of their products.
Thanks to deep learning algorithms, even the smallest defects on heterogeneous or reflective surfaces that are often overlooked by human inspectors can be found with consistently reliable quality, providing powerful insights that help improve quality and operational efficiency.
References
Find out more about the use case.
