Ai-integrated high-speed vision for accurate scrap sorting
This innovative solution combines AI-driven analytics, high-speed 3D machine vision, and robotics to enhance the sorting accuracy of ferrous and non-ferrous scrap metals. It increases throughput and purity while reducing contamination in metal recycling.
Overview
The AI-integrated high-speed vision system offers a cutting-edge solution for improving the sorting accuracy of ferrous and non-ferrous scrap metals. By combining advanced high-speed sensor technologies with AI-driven analytics and traditional magnetic separation, this system enhances the ability to detect subtle differences in metal composition, density, and surface properties. The integration of robotics and machine vision facilitates a comprehensive approach to scrap sorting, significantly increasing throughput, purity, and profitability in metal recycling operations.
Technical specifications
- High-speed 3D machine vision: Utilizes advanced 3D vision techniques for rapid volume and weight estimation, contributing to density calculations and improved metal identification.
- AI-driven analytics: Employs YOLO or R-CNN models with ensemble decision-making algorithms to identify differences in metal alloy composition, geometry, and surface texture.
- Robotic arms with push mechanisms: Enhances sorting speed by using push rather than grasping approaches.
- Gamma correction for image calibration: Refines accuracy by calibrating collected images to detect subtle variances.
- Internal feedback loop: Facilitates real-time process optimization, leading to higher efficiency in sorting operations.
Technology readiness level
Currently at TRL 4, the technology has been developed into a lab-scale prototype, with plans for further validation and scaling to industrial volumes. The project involves partnerships to ensure seamless integration and comprehensive testing, aiming for commercial deployment.
About UC Berkeley
UC Berkeley is a comprehensive public research university of global scale, known for cross‑disciplinary inquiry and a high‑intensity research culture. Industry engages on campus through shared user facilities, project‑based collaborations, and embedded innovation spaces, with proximity to a U.S. Department of Energy national laboratory enabling joint programs and access to specialized instrumentation. Its Bay Area location connects partners to deep talent pipelines and a dense startup ecosystem, enabling rapid prototyping and iteration alongside regional suppliers and investors. Research is supported by competitive federal funding from agencies such as the National Science Foundation, Department of Energy, National Institutes of Health, and DARPA. A dedicated technology transfer office streamlines IP, sponsored research, material transfer agreements, and startup formation, complemented by accelerators and proof‑of‑concept resources.
delivered to your inbox.