| dc.description.abstract | This paper presents a methodology for virtual damage assessment (VDA) of building structures using primarily pre- and post-storm street-level and aerial imagery data. The methodology includes component-based damage assessment on a damage state (DS) scale from no damage (DS0) to complete damage (DS6) for roof, walls, elevated floors, windows and doors, attachments, and foundations, and to estimate the overall damage to the structure. The methodology was applied to assess exterior damage of the 3,408 structures impacted by Hurricane Ian (2022) in Fort Myers Beach, Florida, situated on a barrier island that was completely inundated during the event. The methodology was implemented using engineering students and was validated through a cross-comparison between the assessments performed by the students and a group of ten experts. The cross-validation showed that results obtained by students and experts were within +/− one damage state classification, indicating that the VDA can be conducted by trained engineering students as reliably as experts returning from the field. A second methodology is presented based on previous studies to assess the foundation type and first-floor elevation (FFE) of each building using street-level and aerial imagery. When the foundation type was correctly identified, the FFE estimates had a mean absolute error (MAE) of 1.0 ft (0.31 m), which were significantly more accurate than the FFE estimates from the National Structures Inventory with MAE of 4.27 ft (1.30 m). The resulting data set indicated that distance from the shoreline, building elevation, and year built were strong indicators of overall damage. This data set can be utilized to improve understanding of hurricane damage to buildings and identify important variables, which can lead to the development of reliable models to predict damage in coastal communities. | |
