Increased accuracy of quality inspections through artificial intelligence in the automotive industry

Initial situation

A German automotive company monitored weld seams through worker self-checks
The company operated in a highly regulated market environment with strict quality assurance requirements
Due to insufficient accuracy in the visual inspection system, high rework rates and customer complaints occurred
The company therefore urgently needed to improve the accuracy of its inspection system to prevent impending financial losses

0 %

Time spent on inspection reduced

99,5 %

System accuracy achieved

Causes

Of inaccuracy identified

Introduce an imaging inspection method with AI-supported evaluation
- Generate images of hidden welds using computed tomography (CT)
- Create reference training and test datasets based on defect types and specifications
- Evaluate the performance of different systems based on the state-of-the-art and select the best model
Compile and use high-quality training and test data sets using UMS expertise

Improving assembly processes through optimized development and Six Sigma in the automotive industry

In-line error detection through the introduction of AI-supported inspection systems in the automotive industry

Increased efficiency in quality inspections through AI-supported image recognition in the automotive industry

All cases

Discover more references.