A scalable data collection, characterization, and accounting framework for urban material stocks

Area of Science:

• Built Environment
• Computer Vision
• Circular Economy

Background:

• Building stocks are key secondary resource reservoirs.
• Current characterization methods lack building-specific detail for circular economy strategies.
• A higher-resolution, scalable approach for urban stock characterization is needed.

Purpose of the Study:

• To present a framework for automated, bottom-up characterization of urban building stocks at the component level.
• To enable detailed analysis for informing circular economic strategies.
• To demonstrate a scalable approach for urban stock assessment.

Main Methods:

• Utilizing a mobile-sensing approach combined with computer vision.
• Capturing urban stocks as 3D surface maps for object, component, and material identification.
• Employing a prototype workflow with a custom mobile-sensing platform and neural networks.

Main Results:

• Successfully identified and semantically classified building components (doors, windows) in a Sheffield case study.
• Achieved comparable component counts to manual human counts at both total and building levels.
• Demonstrated the framework's potential for automated estimation of building components.

Conclusions:

• Automated component estimation enhances understanding of circular economy opportunities.
• The framework informs stakeholders for better implementation of circular strategies in refurbishment.
• This approach provides detailed data crucial for advancing the circular economy in the built environment.