Processing and Analysis of Lidar Derived Point Cloud Models of Low-Rise Large-Volume Steel Frame Buildings Damaged in Hurricane Michael

Abstract

Hurricane Michael made landfall in Panama City on the Florida Panhandle on October 10, 2018 as a Category 5 hurricane and caused substantial damage. Low-rise large-volume steel frame buildings (LRLVBs) suffered in particular, and many of them demonstrated similar failure patterns where a short side of the building collapsed, either partially or fully. Following the hurricane, a response team including staff from the NSF RAPID Facility, a natural hazard reconnaissance facility headquartered at the University of Washington, was sent to Panama City where the damage was documented. This was done using lidar scanners, photographs from both handheld cameras and drones, and various other equipment. The use of lidar in reconnaissance to document structural damage following the occurrence of natural hazards is particularly compelling as it can speed data collection relative to physically measuring elements while providing for rich 3D data sets. The accuracy and viability of lidar for use in building reconnaissance was examined in this thesis with a focus on techniques requiring limited computing power. The results show that large scale measurements such as the dimensions of the building, column spacing etc. can be obtained within the margin of error for construction tolerances allowed in AISC steel building construction specifications. It is found that measurements of structural cross-sections are possible from the obtained data but require specific scanner locations and more rigorous post-processing of the data. The results of the work are recommendations for terrestrial lidar reconnaissance specifically related to steel buildings.