Real-time detection and early warning of fire is an important approach to alleviating the threats from fire hazards. Since fire often occurs randomly and a fire scene is usually complicated, traditional fire detection methods are often unable to detect fires and issue warnings in the early stages of fires. Recently, with the abundance of surveillance video cameras, fire detection technology based on video has become an important approach for the early detection of fire. Such detection methods analyze the features of video images and recognize potential occurrences of flames, fires can thus be recognized and under control before they develop into disasters. Due to its ability to detect fires in their early stages, video based fire detection technology has attracted the attention of many researchers in the areas of fire safety and a large number of methods have been developed to detect the occurrence of flames by analyzing sequences of video images.
Color features are the most important features of flames and have been extensively used in methods for video based fire detection. video based flame detection has become an important approach for early detection of fire under complex circumstances. However, the detection accuracy of most existing methods remains unsatisfactory.
Jiaqing Chen and Xiaohui Mu with their team made an research of the flame recognition ,The survey result published on the journal of autonomous intelligence.
In this paper, we develop a new algorithm that can significantly improve the accuracy of flame detection in video images. The algorithm segments a video image and obtains areas that may contain flames by combining a two-step clustering based approach with the RGB color model. A few new dynamic and hierarchical features associated with the suspected regions, including the flicker frequency of flames, are then extracted and analyzed. The algorithm determines whether a suspected region contains flames or not by processing the color and dynamic features of the area altogether with a BP neural network.
Their testing results show that, the approach is robust and able to identify the presence of flames under complex circumstances where other interference sources may also exist. In addition, their approach is able to accurately identify flames that are under control and in safe conditions.
For more information, please visit: