Performance assessment of video-based fire detection system in tunnel environment

摘要

Compared to tunnels used for other modes of transportation, fire safety design problems in road tunnels are more challenging because of high-speed wind environmental impact. A video-based fire detection system (VFDS) applies inventive mathematical calculations and sophisticated computer models to analyze and tackle real-time video signals statistically and intelligently. VFDS-based temporal flicker modeling of flames and wavelet-based contour modeling approaches are used as weak classifiers. Experimental results have shown that false alarms issued by earlier methods can be significantly reduced by using separate flame and non-flame moving pixels. Whenever the system detects fire or smoke under wind or no wind environment, the VFDS server initiates actions, including visual and audible alarms, alarm messages, and video recording.