{"id":5387,"date":"2026-01-08T12:26:27","date_gmt":"2026-01-08T11:26:27","guid":{"rendered":"https:\/\/nl.ai-matters.eu\/getting-started-with-vision-systems-at-ai-matters-meldon-explores-the-power-of-ai-to-improve-processes\/"},"modified":"2026-07-17T16:38:54","modified_gmt":"2026-07-17T14:38:54","slug":"getting-started-with-vision-systems-at-ai-matters-meldon-explores-the-power-of-ai-to-improve-processes","status":"publish","type":"post","link":"https:\/\/nl.ai-matters.eu\/en\/getting-started-with-vision-systems-at-ai-matters-meldon-explores-the-power-of-ai-to-improve-processes\/","title":{"rendered":"Getting Started with Vision Systems at AI-MATTERS: Meldon Explores the Power of AI to Improve Processes"},"content":{"rendered":"\t\t
\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\n

Competition within the manufacturing industry is growing, both in Europe and Asia. Manufacturing companies are therefore constantly looking for ways to work smarter, more efficiently, and with greater precision. Yet for many organizations, AI is still in its infancy. \u201cIf you want to stay relevant, you have to explore the possibilities,\u201d says Stan Erens, COO of Meldon. The company launched an AI-MATTERS experiment in collaboration with Blue Engineering\u2014offered by Brainport Industries Cooperative\u2014to improve product quality through visual inspections. \u201cWe can apply the insights we gained in multiple areas.\u201d <\/p>\n

Meldon: Always on the Move<\/strong><\/p>\n

Meldon develops and manufactures plastic profiles for a wide range of applications. The family-owned company takes a multidisciplinary, innovative approach and works closely with customers and partners. \u201cOur production runs 24\/7, with five shifts and 27 extrusion lines,\u201d says Stan. \u201cWith 95 colleagues, we produce our own PVC compound and granules, and we develop and manufacture our own extrusion tooling.\u201d <\/p>\n

AI does not yet play a major role at Meldon. \u201cWe use robots and have a camera system in two production lines, but those systems are already fifteen years old. They\u2019re due for an upgrade. Through an AI-MATTERS service offered by Brainport Industries Co\u00f6peratie in collaboration with Blue Engineering, we worked with Blue Engineering to investigate which vision technology is suitable for our processes and how we can upgrade our production lines.\u201d <\/p>\n

The challenge: visible defects that are only discovered after the fact<\/p>\n

The experiment focused on a plastic extrusion profile with high visual quality requirements. During the production of these profiles, visible defects\u2014such as stains, scratches, and scuff marks\u2014frequently occur. These defects result in rejection but are often not detected until later in the process. Operators run multiple lines simultaneously and do not continuously monitor the product. With profiles ranging from six to twelve meters in length, it is difficult to trace a detected defect back to the exact moment it occurred. <\/p>\n

Stan: \u201cWe don\u2019t want to sort products after the fact; we want to produce them flawlessly from the start. To achieve that, we need to make product quality measurable and identify deviations during the process.\u201d Machine vision and AI offer a promising foundation for this. They make it possible to collect real-time quality data and optimize process settings in a targeted manner. <\/p>\n

A self-learning system with a 2D line-scan camera and Anomaly Detect<\/strong><\/p>\n

During the AI-MATTERS experiment, Blue Engineering investigated various imaging methods and AI techniques. The tests showed that a 2D line-scan camera is best suited for continuous processes such as the extrusion of these profiles. This camera scans the product line by line and requires minimal lighting, making the system attractive from both a technical and cost perspective. <\/p>\n

At the heart of the quality control process is the AI algorithm Anomaly Detect. This model learns exclusively from images of a good product. The system identifies anything that deviates from this as an anomaly. This aligns perfectly with the variety of potential defects: shape, size, and structure vary significantly depending on the profile type. <\/p>\n

Stan: \u201cMany systems only recognize errors that you define in advance. That takes a lot of time. With Anomaly Detect, it works exactly the opposite way: you teach the model what\u2019s correct, and it automatically recognizes anything that doesn\u2019t fit that pattern. That\u2019s much faster and provides greater reliability.\u201d The system even detects subtle deviations, such as fine scratches. By adjusting the threshold value, engineers can easily fine-tune how sensitively the system responds. <\/p>\n

The feasibility study shows that the combination of a 2D line-scan camera and Anomaly Detect software reliably detects anomalies in various types of profiles and distinguishes them from undamaged products. This represents a significant step toward real-time quality control during production. <\/p>\n

Results and Next Steps<\/strong><\/p>\n

The experiment provided Meldon with valuable insights. \u201cWe now know which technology works for us and how we can use Vision and AI to improve our processes,\u201d says Stan. In the coming period, Meldon will: <\/p>\n