이미지 깊이 추정 기반 Point cloud 및 BIM 모델을 활용한 건축물 피난 및 소방 방해 적치물 모니터링 프레임워크 제안

Proposal of a Framework for Monitoring Building Evacuation and Fire-fighting Obstruction Objects using Point Cloud and BIM Model based on Image Depth Estimation

This study proposes a monitoring system framework for fire compartments and firefighting facilities. Currently, inspections and maintenance of fire safety equipment largely rely on manual, labor-intensive method, making it difficult to achieve real-time monitoring and efficient management through visual inspections alone. As a method, this study presents an image-based obstacle detection methodology using depth estimation within designated fire safety zones. The methodology was applied to a control space for feasibility assessment and further validated on an extended site. The proposed framework was applied to a single floor of H University, where obstacles within stairwells and physical spaces were accurately detected. The integration with BIM models enabled quantitative identification of risk elements through interference analysis. In result, This study presents a lightweight analytical framework that automatically detects obstructions and performs clash checks by aligning point clouds—generated from single-camera images—with BIM models. By selectively modeling only essential monitoring targets for the maintenance phase, the framework offers both practicality and scalability through automated clash detection processes.