Graph-based network generation and CCTV processing techniques for fire evacuation

Evacuation navigation in emergencies such as fires is one of the most important operational considerations for a building. The large and complicated interior spaces, as well as the intensive population significantly increase the difficulty of fire evacuation in large-scale buildings. The environmental changes such as the spread of a fire and the flow of evacuees exacerbate the difficulties of fire evacuation. Therefore, this research aims to develop an adaptive approach for path planning against the rapid environmental changes in fires. In this paper, a graph-based network is formed by integrating MAT with VG, with the addition of a buffer zone. The network uses real-time videos from closed-circuit television (CCTV) cameras facilitated by deep learning algorithms to detect and tally the number of people in a target area. According to the tally of people and a proposed walkability model, the congestion conditions of an area can be analysed so that evacuees can avoid any areas that are congested. An Internet of things sensor network is also established to detect the presence of hazardous areas. The proposed solution allows evacuation navigation to be done in real time. An illustrative example is provided to demonstrate the functionality and features of this proposed methodology.

Automated summary

The paper discusses the importance of evacuation navigation in emergencies like fires and proposes a method that combines real-time video and Internet of Things sensor networks with deep learning algorithms for rapid path planning in changing environments. This approach allows for real-time evacuation navigation in large-scale buildings.