인공지능 기반 전통건축 지붕 기와 손상원인 분류

Abstract

The confirmation of the roof is very important in the maintenance of traditional Korean architecture and architectural cultural heritage, but it is also a place where management is difficult because it is located high. This study photographed traditional Korean architectural roofs using drones, and collected data focusing on "Giwa" among them. The collected data were preprocessed by classifying the cause of damage, and research was conducted to automatically distinguish objects by cause of damage using YOLOv5, an object detection artificial intelligence.