一种基于视频多特征融合的火焰识别算法

根据燃烧过程中火焰的物理特性，结合火焰的静态和动态特征，设计了一种多特征融合的火焰识别算法，对火灾进行快速判别．基于信任度模型建立了火焰识别的概率模型，以疑似概率反映视频图像中出现火焰的几率．实验表明，该算法识别火焰速度快，能够达到25帧／s的处理速度，与其他的火焰识别算法相比，算法在识别过程中无需人工调整即可自动完成，在复杂实际环境中，算法识别的准确性高，抗干扰能力比较强，如对车灯、路灯等干扰疑似概率始终小于10％．因此，算法在实际运用中有着广泛的应用前景．

Flame Detection Algorithm Based on Video Multi-Feature Fusion

The physical flame characteristics in the burning process are analyzed, the intelligent flame detection algorithm is designed with static and dynamic flame features and the probability model of the flame detection is established based on the theory of the degree of belief. The experiments show that the processing rate of the algorithm is as high as 25 frames/s without any manual adjustments. In the complex situations, the algorithm remains with stronger anti-jamming ability and higher accuracy, e. g. , the dubitable probability due to street lamps and automotive lamps can always be lower than 10%.