Occlusion-free Orthophoto Generation for Building Roofs Using UAV Photogrammetric Reconstruction and Digital Twin Data

Building roofs are a primary barrier of the building fabric to protect building interiors and their occupants from the effects of the weather. Effectively monitoring their condition and maintaining them is thus critical, and even more so in light of the increasing risks induced by climate change. Ideally, this requires frequent and careful surveying, as well as recording of the results in a way that facilitates monitoring over time. Traditional manual data collection and defect detection and recording methodologies have been used mainly on building facades. Orthophotos are commonly used for recording. However, (1) current approaches to generate orthophotos cannot ensure the photos are occlusion-free; and (2) the results cannot be efficiently recorded in a structured, digital way with semantic relationship to the corresponding elements in the building Digital Twin (DT). In this paper, a method is proposed to automatically generate high-resolution orthophotos of roof panels from UAV data, given an existing building DT. The resulting orthophotos have high-resolution (e.g. 5mm per pixel). They are occlusion-free and linked semantically to the building elements in the DT. The method is demonstrated using a real case study of a complex pitched slated roof of a traditional building.