Circular Construction Lab

Image

Reusing reclaimed lumber reduces waste, conserves resources, and lowers carbon emissions, supporting more sustainable construction. Effective reuse requires accurate quality assessment to ensure safety and code compliance. Visual grading enhanced by AI-driven image analysis offers a scalable, low-cost alternative to mechanical testing, adaptable to diverse lumber sources while capturing detailed surface characteristics.

Lumber grading traditionally involves measuring mechanical properties and assigning grades under species- and region-specific rules. Conventional Visual Grading depends on trained inspectors to identify defects such as knots or splits, but is often subjective and inconsistent. Machine Stress Rated and Machine Evaluated systems provide greater objectivity through bending tests or sensor-based analysis (e.g., X-ray, ultrasound), yet remain expensive and less suited to irregular reclaimed stock.

Given the variability of reclaimed lumber (irregular dimensions, embedded fasteners, surface weathering, and biological degradation) visual approaches remain the most feasible. AI-assisted grading eliminates subjectivity while leveraging common hardware such as tablets and phones, enabling on-site assessment by deconstruction teams without costly equipment or lab facilities. The system’s algorithms align grading accuracy with standards, reducing inter-grader variation and improving confidence in structural reuse.

AR3-Lumber addresses three challenges: (1) precise 2D image calibration, (2) defect detection across all four longitudinal faces, and (3) auto-grading aligned with existing standards. Integrated image stitching reconstructs full 3D surface geometry for complete defect mapping. Automated grading combines local and aggregated data to generate element-level quality ratings, validated across species and workflows.

By enabling accurate, automated grading of reclaimed lumber, this research supports broader reuse of reclaimed wood in structural applications, builds confidence and capacity for local code enforcers and structural engineers in material and process, enables decentralized local reprocessing and  advances scalable circular construction practices.