Recycleye Ltd has created the world’s largest visual AI dataset of waste materials, working in collaboration with academics from leading European universities. Using this knowledge, the company has developed Recycleye Vision, a database with over 3 million labelled images classifying waste. This system combines AI, smart analytics and low-cost cameras to identify refuse in real time. It can further refine identification by weight and brand detection and is already being used in large materials recovery facilities (MRFs) in England, Northern Ireland, France and Italy.
Recycleye Robotics encompasses a robotic picking system in tandem with the Recycleye Vision database to physically pick up the recyclable waste at an MRF and place it into containers. The robot can pick 55 items a minute and is both more consistently efficient and accurate than human pickers. By applying machine learning, the system is able to detect differences between items such as food and non-food-grade plastics. As a result, the increased accuracy of the process boosts the resale value of the output bales of an MRF’s recyclables by as much as fivefold.
Peter Hedley is the co-founder and CTO at Recycleye, and joined the Shott Scale Up Accelerator programme in 2022. He says, “Our company has expanded so quickly that I am now the manager of other managers! I feel the accelerator scheme will give me the support to improve our planning, culture and technical strategy. The programme will also give me access to a network of other senior SME people who will share valuable knowledge at a critical stage of our growing business.”
Recycleye’s potential has already been recognised and it has raised around £5 million to date. The company will use this funding to scale up and increase the accuracy and capabilities of its machine learning and robotics technology. It will also look to expand its product line and find more European partners.