Rapid Infrastructure Digitization to Support High-Fidelity Hurricane Risk Assessment

Recent losses provide sobering evidence of the continued need for more effective hurricane risk assessment. Fortunately, advances in computational science have now made possible new venues to do so. However, this requires digitization of built environments to allow their evaluation within these computational spaces. Doing so presents a significant challenge, given the diversity and data insufficiency in the structural inventory and the need for an approach that readily scales. In response, this thesis proposes a framework to support automated digitization from publically available images. The framework has three main phases: exterior geometry extraction, interior geometry estimation, and subassembly modeling. Each phase was independently validated and, through the development of a heuristic schema founded in disciplinary perspectives, integrated to demonstrate the resulting fidelity of the approach. The end result is a digitization schema that can be readily applied to wood frame residential structures and is readily expandable to other construction typologies.