Case Study: Glacier Robotics reduces PET leakage with Ultralytics YOLO11

Glacier Robotics reduces PET leakage by 70% with Ultralytics

Glacier Robotics, a company that builds AI-powered robotic sorting systems and analytics tools for recycling facilities, faced the challenge of accurately identifying and classifying a wide variety of overlapping, dirty, and visually heterogeneous materials on fast-moving conveyor belts. To improve performance and scale their deployments across multiple facilities, they needed a more rigorous and accurate computer vision model than their previous solution. They turned to Ultralytics and implemented the YOLO11 model.

By integrating Ultralytics YOLO11 into both their robotic sorting arms and cloud-based analytics platform, Glacier achieved significant improvements in classification accuracy and operational visibility. The solution enabled real-time object detection for robotic picking and provided valuable data insights for facility operators. The results were substantial, including a 70% reduction in PET bottle leakage at one facility, which recovered 15 million bottles and generated $138,000 in new revenue. Robots also achieved 95% uptime, demonstrating the reliability of the Ultralytics-powered system.

Glacier Robotics