A novel way for fire detection in the video using Hidden Markov Model

Computer vision and pattern recognition techniques have been recently applied to the fire detection. Unfortunately the complex environmental condition in some scenarios presents a unique challenge to the existing algorithms. In this paper, we propose a hierarchical framework for a more efficient fire detection. Firstly, we follow some classic algorithms to find the "Candidate Area" (CA) on each original video frame. Then, a group of "Confident Points" (CPs) are arranged around the boundary of a CA, after which a modified Hidden Markov Model (HMM) is used to describe the fire\´s flickering characteristic. Final experiments demonstrate that our proposed algorithms can well keep the robustness and stand the complexity imposed by varied outdoor circumstances, compared with previous algorithms.