Video Fire Detection (VFD) is one of the most active research topics being valuable for both theoretical and practical research in computer vision

especially has a wide spectrum of promising applications in video surveillance for early fire alarms in public security. As the improvement on visual feature model of fire

many VFD systems have been developed. In this paper

some main issues on VFD are reviewed

including its advantages to traditional detectors

the classification and description for visual fire features

the representative algorithms and systems

the future trends

and so on. Then some key problems on the compatibility

real time efficiency

intelligence

performance evaluation and multi sensor fusion for VFD are discussed. In addition

a novel VFD model based on hierarchical attention and a saliency fusion framework based on multi sensors are proposed for boosting the efficiency and activity of fire surveillance by using salient feature representation and low computational redundancy.