科研之友
微信
新浪微博
Facebook
分享链接
摘要
Nowadays early fire detection and extinction becomes imperative to prevent large fires and protect life and goods. The multi-sensor fusion based fire detection and fighting system is presented in this paper to improve resistance to nuisance alarm. A cascaded two level fusion scheme is performed to calculate fire distribution probability. A fire alarm is triggered whenever the fire probability is above the relevant threshold. In the first level, the fire probability is inferred by a machine vision sensor in order to decrease the time delay of the fire detection process, whereas the second level utilizes in-field sensors to deal with false alarm induced by vision sensors. The machine vision sensor uses the motion history detection algorithm to register the possible fire position in transmitted video data and then analyze the spectral, spatial and temporal characteristics of the fire regions in the image sequences. An intelligent fire fighting robot equipped with a non-contact in-field temperature sensor and an ultraviolet radiation sensor is guided by visual servoing feedback framework to accomplish adequate reliable fire detection task and reject nuisance alarm. Experimental results demonstrate that the sensor fusion based intelligent fire fighting robot can successfully detect the fire incidents and discriminate nuisance sources reliably.
- 出版日期2011-7
