一种基于视频的隧道火灾检测算法

A video-based tunnel fire detection algorithm

交通视频处理中隧道火灾检测是一个重要的主题。本文提出了一种基于视频的隧道火灾检测新方法。该方法依据火焰及烟雾的特征,先找出前景中火焰烟雾区域,再利用BP神经网络和D_S证据理论进行数据融合实现火灾检测。实验结果表明,该方法实时性好、准确性强,具有很大的实用价值。

Tunnel fire detection is an important topic in traffic video processing.This paper proposes a new method of videobased tunnel fire detection method. According to the characteristics of flame and smoke, the method firstly find out flame and smoke region in the foreground ,then use BP neural network theory and D_S evidence theory for data fusion so that we can achieve fire detection.The tests show that the method is real-time and accurate, so it has great practical value.