Recycleye Vision – World-Beating Machine Learning Accuracy
Combining AI, smart analytics, and low-cost cameras, Recycleye Vision is an end-to-end operating system that provides facilities with unprecedented visibility on waste composition, and operations. While current UK regulation requires facilities to sample just 0.05% of waste, Recycleye Vision analyses 100% of it, with computer vision the only technological breakthrough that can achieve such visibility on waste streams without packaging or production processes being physically altered.
Recycleye Vision is underpinned by WasteNet, the world’s largest visual database of waste items, developed in partnership with Imperial College London and the Delft University of Technology in the Netherlands.
Consisting of over 2,700,000 labelled images (with this number growing continuously), WasteNet uses intelligent software and computer vision algorithms to replicate, and ultimately outperform, human vision, and identify waste items at a brand and weight level. By applying machine learning, Recycleye is able to increase the granularity of waste sorting by, for example, detecting food vs non-food-grade plastics (which to-date has not been possible), and in the process increase the resale value of bales fivefold. The data is for the first time commoditising waste by providing a detailed digital breakdown.
Installed within more than 10 facilities across the UK, France, and Italy, Recycleye Vision also allows recycling facilities to gain full visibility of their operations, providing high-level and actionable insights, and metrics on stream purity, plant downtime, hazards, and potential plant failures. For example, it will detect if an item is potentially large enough to cause a piece of machinery to break.
At Biffa’s Aldridge facility, Recycleye Vision has been installed to automate the quality assessment process entirely, classifying individual items on their PP, PET and HDPE lines. Mick Davis, Chief Operating Officer for the Recourses and Energy Division at Biffa, comments: “It is great to have the opportunity to partner with Recycleye at our Aldridge facility. The technology gives us a much better picture of the waste we are collecting as well as optimising the processes at the plant. From spotting hazardous material which could damage machines to identifying metrics which can improve working patterns, the AI technology can help transform the way that our waste is screened and sorted. This is not only a positive step forward for Biffa, our employees and customers, but better for the environment and the Government’s increased recycling rate target.”
