主要研究了火焰的物理特征信息, 结合图像处理技术, 提出了一种基于图像中高亮度区域特征的森林烟火识别方法。该方法的核心是采用分步提取预处理后图像颜色、面积及圆形度特征来实现烟火识别, 弥补了单个算法仅对单一特征进行分析导致识别准确度不高的缺点, 且能快速排除干扰物, 提升了整个森林火灾图像识别的运行效率。

This paper mainly studies the physical characteristic information of the flame. Combined with image processing technology, a kind of forest fireworks recognition method based on the characteristics of high brightness regional image is proposed. The key of the method is that it uses the feature of pretreatment image color, size and roundness to achieve fireworks recognition. It makes up the disadvantage of single algorithm with low accuracy which only analyzes single feature. Also, it can quickly eliminate the interference, improve the operational efficiency of the whole forest fireworks image recognition.