Siemens and P&G Deploy AI Quality Inspection System Globally, Reducing Scrap Rates by 10% to 20%

en.Wedoany.com Reported - On September 16, Siemens and P&G announced the expansion of the deployment of the Visual Inspection Cockpit (VIC) AI-powered visual quality inspection system, rolling it out across P&G's global manufacturing network. VIC can continuously inspect products while production lines run at full speed. Depending on the product and production conditions, existing deployments have reduced scrap rates by 10% to 20%, and new production line system commissioning is 5 to 10 times faster than traditional custom machine vision solutions.

VIC was jointly developed by Siemens and P&G, combining P&G's deep learning models with Siemens' Industrial Edge edge computing platform, and employing industrial computers configured with NVIDIA GPUs along with supporting AI software and hardware. Image data is processed directly near the production equipment, eliminating the need to transmit all inspection tasks to the cloud, and inspection results can be fed into the production control system in real time.

The system is designed for high-speed consumer goods manufacturing lines, with a focus on flexible, textured, and highly variable products and packaging materials. During high-speed operation, such materials may develop defects such as movement, stretching, wrinkling, overlapping, and low-contrast defects. VIC combines AI semantic segmentation with traditional machine vision logic, updating inspection models when materials, packaging designs, and production environments change, while maintaining continuous online inspection.

VIC can identify overlapping components, low-contrast defects, and abnormal conditions in intricately decorated products and packaging, while connecting to PLC control systems in real time. Upon detecting non-conforming products, the system can directly trigger alarms or perform individual item rejection, and continuously writes quality data to the manufacturing system for production trend analysis and process adjustment.

Previously, P&G needed to configure visual inspection systems separately for different products, packaging formats, and production line environments. The new solution standardizes VIC and its accompanying Engineering Tool as Industrial Edge applications. New product inspection tasks can be deployed to more production lines through model updates, reducing the engineering work of rewriting dedicated vision programs and reconfiguring inspection systems.

Siemens disclosed that the system can operate at production speeds of thousands of products per minute and scale from a single production line to multi-factory deployment. Its industrial AI vision solution can process up to approximately 2,400 images per minute in relevant applications and feed inspection results back to automation systems in real time.

The global deployment will continue to use Siemens Industrial Edge as the unified industrial computing infrastructure. Siemens is responsible for the industrial computing platform, software expansion, and long-term operational support, while P&G is responsible for product inspection models and manufacturing applications. The system has progressed from single-line AI visual inspection to cross-factory replicated deployment.