Fraunhofer IIS and Lobbe develop innovative AI X-ray sorting system
Greater safety in recycling plants 13.01.2026Detecting hazards early, preventing fires, strengthening cycles: With DangerSort, the Fraunhofer Institute for Integrated Circuits IIS and the Lobbe Group have developed an innovative AI X-ray sorting system. Lithium-ion batteries are reliably detected in packaging waste and automatically ejected. The jointly developed technology is now in operation at Lobbe in Iserlohn.
© Lobbe Group
Lithium-ion batteries are among the greatest fire hazards in sorting and recycling facilities. Thousands of fires occur every year – most of them caused by improperly disposed batteries in electrical appliances. In order to significantly improve the protection of facilities, people, and infrastructure, Fraunhofer IIS and Lobbe have spent the last three years developing and extensively testing the AI-supported X-ray sorting system DangerSort.
The system has recently been running in test mode at Lobbe. During the sorting process, an X-ray source scans the packaging waste passing by on the high-speed conveyor belt – similar to baggage screening at an airport. A detector continuously captures X-ray images, which are analyzed in real time by specially trained AI. If the system detects a hidden or concealed battery, precision nozzles trigger air blasts in milliseconds, which eject the object into a separate container. This reliably removes hazardous items from the material flow without affecting the throughput of the system.
In practical tests, DangerSort identified a strikingly high proportion of incorrectly disposed-of waste electrical equipment and achieved a yield rate of over 90 %. Even larger energy storage devices, such as tool batteries, could be reliably separated – incorrect disposals remained the exception. "For us as plant operators, this system represents a quantum leap in terms of safety," says Jochen Ebbing, process engineer at Lobbe. "The technology significantly increases fire safety while protecting our employees and equipment."
The commissioning is also an important step for Fraunhofer IIS: "The combination of X-ray technology and AI enables reliable detection of batteries under real-world conditions," explains group manager Johannes Leisner. "Continuous retraining of the AI further increases the robustness of the system."
DangerSort is part of the K3I-Cycling project funded by the German Federal Ministry of Research, Technology, and Space and is anchored in the AI application hub for plastic packaging. In this alliance, partners from industry, science, and society are working to make AI usable for a more efficient and even safer circular economy. The collaboration between Fraunhofer and Lobbe makes it possible to evaluate and further develop the technology directly in an industrial environment.
