vAI from Viscom – Enhancing Performance with Artificial Intelligence

In automated quality assurance, the distinction between "acceptable component" and "defective component" can be so fine that the results need to be reclassified. Viscom's AI based verification leverages deep learning to assist production sites. Following the four-eyes principle, the AI's results are compared with the operator's decisions. If there's a deviation, a notification prompts for correction or confirmation of the input, preventing misclassification early on. Transitioning smoothly to autonomous classification, the AI can make independent decisions in predefined areas, significantly reducing time.

When traditional inspection algorithms hit their limits – such as in the fast and accurate localization of voids in thermal connections, battery cell anodes and cathodes, or wire bond paths – AI offers compelling advantages. Even with heavily distorted image data, it delivers perfectly segmented results. This enhances the reliability of actual error detection and improves the First Pass Yield. A stable AI model is trained in advance and directly implemented in the inspection plan.

When all the necessary information is available, an inspection plan for electronic components quality control can be quickly created using conventional methods. However, if data is missing or incomplete, AI can identify components and solder joints. The AI then automatically assigns the appropriate inspection patterns. This approach allows AI supported inspection plan creation to fully utilize its advantages in both prototypes and high-mix production.