Advanced Computer Vision Frameworks for Automated Construction and Demolition Waste Valorisation

Sorting construction and demolition waste for recycling is currently a slow, dangerous, and manual process. This thesis developed advanced computer vision systems to automate this task. By teaching computers to "see" and understand complex, messy piles of waste in real-time using live 3D camera data, this research created the software necessary to guide robotic sorting arms. This technology provides a high-speed, automated solution to safely recover valuable building materials like wood, metal, and concrete, helping the construction industry reduce landfill waste and transition toward a sustainable circular economy.