실내 지능형 대피 시스템을 위한 피난안내도 기반 실내공간정보 구축

Indoor disasters occur frequently, leading to significant loss of life and property. As indoor spaces are where people spend the majority of their daily lives, disasters in these environments carry a high risk of resulting in substantial casualties. To mitigate such risks, various studies have explored digital-based intelligent evacuation systems that aim to minimize damage and support efficient evacuation during emergencies. Spatial information plays a critical role in the implementation of such systems, however, the high costs associated with constructing spatial datasets present a significant challenge. Evacuation maps, which are typically mandated by law in most buildings, provide a readily available and cost-effective source of indoor spatial information. By leveraging these maps, it is possible to construct spatial datasets for indoor environments more efficiently and affordably. In this study, we aim to detect and extract information from evacuation maps to create structured data that can be directly utilized in intelligent evacuation systems. Specifically, we propose a method for automatically identifying wall structures and semantic information about indoor spaces from evacuation maps. Furthermore, we suggest methodologies for organizing and storing this information and discuss its potential applications in intelligent evacuation systems and beyond.