一种基于图像多特征的火焰识别算法

火灾探测技术的稳定性高、误判率低是目前研究的热点,重点分析了火灾火焰燃烧过程中色彩形态的变化过程.提出用火焰色彩、面积变化、尖角数、质心位置移动等信息作为火灾识别的依据,采用逐步排除、渐进的识别方式作出火灾判断的算法.实验结果表明,基于图像多特征的火焰识别算法能够有效识别出火灾火焰,并有很强的抗干扰能力.

Flame Detection Algorithm Based on Multi-feature of Image

High stability and low misjudgment rate for fire detection technology is a hotspot of cur-rent research. This paper analyzes the changing process of color and shape in the process of combustion. Four criteria have been mentioned ： the color of flame, the area changes gradually, the numberof sharp comers, and the moving center of mass. Stepwise and progressive means of identificationwere used to judge the fire. Experiments show that flame detection algorithm based on multi - feature of image can recognize the fire flame effectively and have strong anti - interference ability.