Computer-Vision Based Aluminum Scrap Grade Detection for Circular Upcycling

Technology
In development
University
Scouted & summarized by Halo AI

An AI-driven computer vision system that identifies and sorts post-shredding aluminum scrap by grade with 98.4% accuracy. By analyzing unique shredding patterns and surface textures, the technology enables high-value material recovery, reducing waste and supporting sustainable wrought-to-wrought recycling loops.

Overview

The University of Warwick has developed an advanced computer vision-based classification algorithm designed to identify commercial grades of post-shredding aluminum scrap with 98.4% accuracy. By automating the identification of scrap metal, this technology enables efficient sorting back into high-value wrought-to-wrought recycling loops. This software-driven solution leverages the micro-structural and surface traces—such as texture and cutting marks—created during the shredding process to distinguish between different alloy classes, supporting the transition toward a more sustainable circular economy.

Technical specifications
  • Capability: Automated classification of post-shredding commercial aluminum scrap into specific alloy grades.
  • Accuracy: 98.4% performance verified in identifying shredded scrap materials.
  • Methodology: Utilizes deep learning neural network architectures, including Vision Transformers (ViT), to analyze RGB images captured by standard, off-the-shelf digital cameras.
  • Feature Extraction: The system identifies discriminative features such as material folds, edge characteristics, and surface textures unique to the shredding process.
  • Hardware Accessibility: Designed for compatibility with cost-effective industrial camera setups, reducing the need for specialized, expensive sensor equipment.
Technology readiness level

The technology is currently at a maturity level of 6. The research has successfully demonstrated high-accuracy classification (98.4%) for shredded scrap using curated industrial datasets and peer-reviewed deep learning methods. It has progressed beyond basic lab validation to demonstrate efficacy in industrial contexts, though it is categorized as an established research outcome ready for further pilot integration and commercial development.

Generated Jul 8, 2026. Please verify key information.

About University of Warwick

The University of Warwick is a comprehensive public research university in Coventry, UK, with a strong applied innovation culture. Industry collaboration is embedded through WMG, a High Value Manufacturing Catapult centre that provides co‑located labs, pilot‑scale facilities, and industrially sponsored programs tightly integrated with company teams. On campus, the National Automotive Innovation Centre—created with Jaguar Land Rover and Tata Motors—offers shared R&D space for rapid prototyping and joint development. A university‑owned science park and flexible collaboration mechanisms, including contract research and Knowledge Transfer Partnerships, connect corporates and SMEs across the Midlands ecosystem. Research is backed by UKRI councils and Innovate UK, while Warwick Innovations manages IP, licensing, and startup formation.

Halo home
Partner smarter. Move faster.
Get new partnering requests
delivered to your inbox.