AI meets cheese production: rethinking efficiency

Care robot, 4K line scan camera and deep learning: how Eberle automatically inspects every cheese wheel for Gebr. Baldauf, reduces waste and secures quality.

Care robot with cheese wheels and a blue camera housing in the cheese cellar

The food industry benefits greatly from advances in artificial intelligence. The combination of AI with machine vision is particularly valuable: it enables entirely new applications and increases productivity, especially in quality assurance. One example is a fully automated system of a mobile care robot and image processing that reduces the waste of cheese wheels during maturing, improves quality and cuts costs.

Challenges in the cheese industry

Cheese is popular worldwide and demand is rising. At the same time, European cheese dairies in particular are feeling the skills shortage, sustainability is becoming more important, and the industry wants to reduce waste to conserve resources. On top of this, consumers expect more quality and variety.

“One solution to these challenges is the end-to-end automation of the maturing process,” says Dorian Köpfle, Machine Vision Engineer at Eberle. “Cheese wheels often mature in climate-controlled rooms for up to 14 months. High storage costs and the wish for minimal waste require regular checks for mould. But inspecting thousands of wheels by hand is impossible. That is why the Gebr. Baldauf cheese dairy commissioned us to develop an automated solution.”

Gebr. Baldauf GmbH & Co. KG is a long-established cheese dairy in the Allgäu region. Eberle specialises in automation solutions for the food industry, robotics and industrial image processing. Eberle developed the solution together with Leu-Anlagenbau, a Swiss company that supplied the mobile care robot.

A fully automated solution

The goal was complete monitoring throughout the entire maturing period, which had not been possible by hand. This reduces waste because deviations are detected early, and improves quality through individual care of each wheel. Added to this are standardised quality control, full traceability and greater efficiency. The system also forms the basis for long-term data analysis and future AI applications.

  • Care robot with bar lighting and a 4K colour line scan camera
  • compact industrial PC on the robot, central computer at the customer for deep learning classification
  • MVTec HALCON for image processing, Eberle’s storage for monitoring and documenting the care process
Camera images of the inspected cheese wheels, defects are marked in a heatmap

AI and deep learning in practice

The robot moves through the cheese cellar and tends the wheels in groups of three, each lying on a board. One image is taken per group, pre-processed and sent to the central computer, where it is classified using deep learning. Position data determine the type of cheese and a unique cheese ID. Data and analysis are accessible via the warehouse management system.

The cameras in the blue housing capture images of the cheese wheels inside the care robot. In the visualisation of the inspected cheese wheels, defects are highlighted using a heatmap.

Technical challenges

The main technical challenge was the variety of cheese types and stages of maturity, which look very different and change considerably during maturing. Classic image processing reaches its limits here, so HALCON detects anomalies using deep learning.

“HALCON reliably detects anomalies of any kind by combining different methods. First the relevant image areas are identified, then classification and analysis follow,” explains Felix Podhorsky, Business Development Manager at MVTec. Christoph Muxel, responsible for customer relations at Eberle, adds: “We have been working with MVTec for several years because we are convinced by the range of functions, the performance and the ease of use, even with a wide variety of cameras.”

For the inspection to run fully automatically, a training model was needed that takes the differences between types and stages of maturity into account. Today the customer can carry out the classification via the warehouse management system; Eberle is working towards fully automatic inspection.

Goals achieved

The system has been in operation since December 2024 and has exceeded expectations. “Our cheese quality control is a flagship project for digitalisation and automation in the food industry. The customer has been able to reduce costs, improve quality assurance and lay the foundation for further smart processes,” says Dorian Köpfle.

Manual work decreased while the frequency of inspections increased. Because cracks or mould are detected early, waste went down. The system provides complete digital documentation, automated quality checks and end-to-end traceability. In the long term, in-depth data analysis will help to further optimise the maturing process. Furthermore, the system is ready for integration with ERP or cloud platforms and can be extended to further types of cheese.

Following the success of the pilot project, Eberle plans to develop the system into a standard solution for the cheese industry. “We want to integrate the camera system into mobile care robots and stationary care systems worldwide. The project in the Allgäu region shows how machine vision can help to enhance quality, efficiency and competitiveness in cheese production,” says Christoph Muxel. More on the technology under industrial image processing.

Interested?

Talk to our experts about your project.

  • Free & non-binding
  • Individual advice from experts
  • Solutions from special machine building practice
  • 30 minutes — no small talk
Viktor Eberle
Geschäftsführung