State of the art in vision-based fire and smoke dectection

Abstract

Video processing techniques for automatic fire and smoke detection have become a hot topic in computer vision during the last decade.The several vision-based detection algorithms that have been proposed in literature have lead to a large amount of techniques that can be used to detect the presence of fire at an early stage.In this paper, we perform a thorough evaluation of the existing techniques and propose an improved smoke detection algorithm.The novel algorithm is based on a chromaticity-based background subtraction method with back-step correction, wavelet-based energy analysis, and boundary disorder analysis.These steps are chosen because they are mainly independent of the burning material and the fire circumstances.Experiments with a large number of fire and non-fire video sequences show that our method is fast and capable of accurately detecting smoke.