The Low-Power Fire Recognition Algorithm under Complex Background
According to the fire burning features, this paper proposes a low-power fire detection algorithm under complex background. In this algorithm, adaptive statistical background features are used as the basis for background suppression to remove the background image effectively and local accumulation area method is used to eliminate interference and identify areas of fire flames. The algorithm solves a difficult problem that low-power fire detection in video sequences which exists in complex background and low SNR is difficult to get. The experimentation results show that under the condition that the image resolution is 400 × 300, the algorithm proposed in this paper can identify the flame within 100 pixels with algorithm response time less than 7s, recognition rate more than 90% and false positive rate less than 10%.
Q. Liu et al., "The Low-Power Fire Recognition Algorithm under Complex Background", Applied Mechanics and Materials, Vols. 58-60, pp. 501-506, 2011