👁️‍🗨️ Revolutionizing Manual Assembly with Machine Vision: Endress+Hauser's Success in Flow Meter Production!

[Elif Özaksu]

🚀 Endress+Hauser's Smart Manufacturing Move​

Endress+Hauser Flow has taken a revolutionary step in high-variation, low-volume flow meter assembly processes. In this area where full automation is impractical, they have both improved quality and increased efficiency with a machine vision-supported manual assembly system.

🔍 A New Era in Quality Control​

Quality control processes, previously relying on the supervision of two employees, are now automated thanks to MVTec MERLIC software. This integration ensures early detection of errors, guaranteeing a smooth transition of products to subsequent production stages. The company demonstrated its sensitivity to data security and accessibility by opting for a local system integrated with its existing Manufacturing Execution System (MES) instead of cloud-based solutions.

⚙️ Seamless Integration with MERLIC​

MVTec MERLIC has been seamlessly integrated into Endress+Hauser's web-based MES. Before assembly, the system verifies the correct process step, the appropriate workstation, and the authorized employee. Then, assembly instructions enriched with text and images appear on the screen. MERLIC works in the background, analyzing images taken during assembly and confirming that critical operations are performed correctly. All inspection results, process information, and image data are recorded by the MES.

💡 Human and Machine Collaboration​

This system complements human labor rather than eliminating it. While high product variety is managed through human assembly, image-based inspection automatically verifies specific process steps. Thus, the flexibility of manual labor is combined with the precision of machine vision.

🎯 Verification of Critical Steps​

A camera mounted on a height-adjustable assembly table observes the operations performed on the workpiece. Barcode readers determine the specific assembly sequence and requirements for each component. MERLIC provides additional quality assurance by checking critical steps, such as whether gaskets are installed correctly. An industrial computer, a camera with a variable-focus liquid lens, and a laser distance sensor form the core components of this system.

🧠 AI-Powered Image Analysis​

MERLIC's two AI-based technologies – semantic segmentation and object detection – are used for image-based recognition. These technologies localize error classes at the pixel level and identify objects with bounding boxes. When a process is completed correctly, the system automatically moves to the next assembly instruction, reducing the need for manual approval.

🌍 Global Rollout and Future Plans​

Endress+Hauser has been using this system since March 2025 and began rolling it out to other global production facilities in autumn 2025 after the initial deployment. In the future, MERLIC aims to increase the level of automatic identification by reading component codes with barcode reading and optical character recognition (OCR). This will allow the employee to initiate the entire assembly and inspection process simply by holding the workpiece under the camera.

This hybrid approach demonstrates how machine vision can be effectively used in manual production environments. By combining the flexibility of human assembly with the consistency of automated inspection, Endress+Hauser is both improving quality and optimizing production processes.