Title Artificial Intelligence-Based Detection of Damaged Roof Tile for the Maintenance of Traditional Buildings
Authors 임수영(Lim, Su Young) ; 전한종(Jun, Han-Jong)
DOI https://doi.org/10.5659/JAIK.2023.39.6.39
Page pp.39-46
ISSN 2733-6247
Keywords Traditional architecture; Roof tiles; Building maintenance; artificial intelligence; drone
Abstract The roof of traditional Korean buildings is a part that suffers much damage, whereas monitoring is not performed smoothly due to accessibility difficulties. This study aims to detect automatically damaged roof tiles using AI and drones to maintain traditional buildings. Information on roof tiles is extracted through drone image photography, and an automatic object detection function is utilized through the YOLOv5 algorithm. As a result, through this study on the maintenance of traditional Korean buildings, it is possible to easily inspect the tiles and check the damage status of the roof tiles.