AI in Industrial Production
VisionAI: Enabling Customers to Train Their Own AI
At their Open House 2026 in Lübbecke, IMA Schelling is presenting VisionAI for the first time: an AI platform that allows customers to train their own models for industrial image analysis. In wood processing, initial applications for production control and quality inspection are already in use with reference customers. VisionAI expands IMA Schelling’s existing vision portfolio and opens it up to use cases defined by the customer.
VisionAI Expands Existing Vision Portfolio
At IMA Schelling, image processing and inspection tasks are now handled by three specialized optical systems. Optonic detects positions and gripping points in robot-assisted panel picking. Fagus-GreCon uses laser-based geometry measurement to check edge quality after edge banding. Both systems perform their tasks precisely and can be seen at the Open House in Lübbecke. While Optonic and Fagus-GreCon handle clearly defined tasks, VisionAI opens the field to inspection tasks defined by the customer. This creates numerous new use cases, including applications beyond the machine itself.
VisionAI Cloud: Custom Models for Individual Use Cases
Through a license-based access to the VisionAI Cloud, customers can train their models directly using their own production data – without complex programming and without external specialists. The AI learns which features in the production process correspond to a target state and can, for example, report deviations or frequencies. In this way, an unlimited number of process steps can be monitored and certified – even in storage or goods receipt.
The cloud platform itself is hardware-agnostic. Existing cameras can be used, and additional compatible hardware can be freely selected or purchased from IMA Schelling. Once trained, AI models can be easily expanded with new applications or additional inspection features. “We want to enable our customers to set up their own use cases in production and gradually discover the wide range of possibilities offered by this technology,” says Hasan Akpinar, Business Operations Manager Vision AI & Service Consulting at IMA Schelling.
Anwendungen für die Holzverarbeitung
Neben der Cloud-Plattform gibt es bereits konkrete Anwendungen für die Möbelindustrie, die bei ausgewählten Referenzkunden im Einsatz sind. Die automatische Bohrlocherkennung erkennt zum Beispiel Werkzeugbruch, falsche Bestückung oder Bedienfehler während des laufenden Prozesses. Die individuell definierbare Oberflächenerkennung ersetzt manuelle Stichproben durch eine KI-basierte Bewertung in Echtzeit. Fehlerhafte Teile schleust das System aus, noch bevor sie weiterbearbeitet werden. Dadurch sinkt der händische Aufwand bei der Qualitätskontrolle. Gleichzeitig dokumentiert das System den gesamten Prozess und die Daten lassen sich anschließend weiterverarbeiten.
Applications for Wood Processing
IMA Schelling’s integration experts ensure where potential integration points for VisionAI exist within the system and how the results feed back into control systems. “The VisionAI Cloud alone does not provide real added value. What matters is what you focus on in the production process and how the results improve efficiency. The possibilities are diverse,” says Hasan Akpinar. The AI architecture is provided by the technology partner Denkweit from Halle, a Fraunhofer spin-off with seven years of experience in AI-based image recognition. IMA Schelling is responsible for industrial expertise, global sales, and system integration.
From Pilot Project to Use-Case-Raffle
VisionAI has been in an extensive pilot phase for some time. Currently, IMA Schelling is testing drill hole detection in live operation at industrial cycle speeds. The system compares defined target and actual states and automatically removes defective parts. Initial results show that AI applications on machines reduce production errors and increase economic efficiency. From May 6 to 8, 2026, IMA Schelling will showcase VisionAI at the Open House in Lübbecke. Visitors can submit their own production use cases in a Use-Case-Raffle. The most compelling use case will be implemented, and the customer will receive a free VisionAI Cloud license for 12 months.
Outlook: Beyond Wood Processing
Since customers train the models themselves, VisionAI is not limited to wood processing. One example is detecting cut surfaces in the metal industry. Cutting errors waste material and tools; detecting them early keeps processes running and prevents costly rework. The solution can also be used beyond machines, for example for item counting in goods receipt. VisionAI is now available from IMA Schelling in various licensing models.