10182173Computer vision–driven multisensor verification for resource-efficient Recycling: Integrating RGB, NIR, and Thermal Imaging to identify contaminants in recycled metalsClosedCollaborative R&DInnovate UKAgave Networks will carry out a two-month feasibility project to explore how computer vision and multisensor imaging can improve the quality and resource efficiency of recycled-metal trading. The company already develops digital tools that help recycling businesses record and verify container-loading operations. Building on this experience, the project will test a small prototype that combines standard colour (RGB) cameras, near-infrared (NIR) sensors, and thermal imaging to identify contaminants and quality issues in scrap-metal loads. Today, metal-recycling yards depend mainly on human visual inspection and trust between buyers and sellers. This manual approach can lead to errors, material downgrades, and unnecessary waste. The proposed system introduces a low-cost, data-driven method that can highlight non-metallic materials such as plastics, rubber, or wood and flag unusual heat patterns that might indicate contamination or safety risks. By giving operators a clear, objective picture of each load, the technology aims to support cleaner scrap, fewer rejections, and better use of existing resources. The study will take place in two stages. First, the team will run controlled tests using known materials in a laboratory environment to confirm that the NIR and thermal sensors can distinguish metals from non-metals. Then, the same setup will be tested at a UK scrap-metal yard to gather real-world images and feedback from industry users. Simple computer-vision algorithms will process the images to compare results from RGB-only cameras with those from the new multisensor approach. The outputs will include demonstration footage, contamination maps, and a short report summarising technical performance and environmental benefits. This project is innovative because it brings together accessible hardware and practical computer-vision techniques to address a long-standing challenge in recycling: the lack of reliable, automated quality verification. The approach does not require complex machine-learning models or expensive laboratory equipment, making it suitable for small and medium-sized recyclers as well as larger operators. By proving the concept at low cost, Agave Networks will create the foundation for a larger programme of work focused on AI-driven material verification and digital traceability. The results will contribute to the UK's goals for net-zero manufacturing, circular-economy growth, and resilient material supply chains, showing how digital innovation can make recycling cleaner, safer, and more efficient.