Future directions for video fire detection

摘要: To accomplish more valuable and accurate video fire detection, this paper points out future directions discusses first steps which are now being taken to improve the vision-based detection of smoke flames. First, an overview is given state-of-the-art methods in visible infrared spectral range. Then, a novel multi-sensor flame detector proposed combines multimodal information low-cost visual thermal results. Experiments on nonfire sequences indicate that combined yields results, with fewer false alarms, than either alone. Next, framework for multi-view analysis discussed overcome lack video-based tool detect characteristics at early stage fire. As prior experimental results show, from different viewpoints provides characteristics. Information about 3-D location, size growth rate can be extracted data practically no time. Finally, towards standardized evaluation video-driven forecasting suggested.