A survey on video smoke detection

Abstract

Fire destroys human lives and property. Therefore, there is a huge need for a reliable and probable fire detection technique. This paper provides a review on various methods developed to detect smoke through videos. The study basically categorizes techniques of smoke detection on the basis of feature extraction method (static/dynamic characteristics), locating region of interest (ROI), etc. It also discusses the nature of camera, color model used for detection and so on. A basic method of smoke detection is described stepwise with different types of algorithms used in each step. The pros and cons of each method are also discussed briefly in this paper.