This automotive customer developed a body shop that used manually processes to inspect stamped parts to detect common defects such as splits, strains and hole counts. Multiple operators being used and the number of defected parts passing through was leading to huge costs. The objective was to reduce the number of defected parts passing through the line by having a system to support the operator and increase control measures. 

This lead to an AI solution from Siemens digitilising this process. The line itself being designed on Process Simulate to identify the most effective positions for sensors/cameras as well as producing synthetic data to train the AI model. The physical commissioning included an edge device to collect the shop floor data that passes through vision quality applications and inferred in the AI model with an output visualised using a HMI. This shopfloor data is then reused for fine tuning the model to increase reliability of the system.