摘要
结合火焰静态和动态特征,设计一种应用于变电站的火焰识别融合算法。将静态特征划分为颜色、形状和纹理特征,将动态特征划分为区域增长特性和闪烁特性,利用先进的网络摄像机技术,抓拍图像或视频,取静态/动态检测结果交集,实现火焰的识别。搭建软硬件环境开展实验,表明火焰识别融合算法在正常光照下的识别率大于99%,且对复杂背景、高强光环境具有较强的适应性。
Combining static and dynamic characteristics of flames, an integrated flame recognition algorithm applied to substations is designed. The static features are divided into color, shape and texture, and the dynamic features are divided into regional growth characteristics and flicker characteristics. By using advanced network camera technology, images or videos are captured, and the static and dynamic detection results are integrated to realize the flame recognition. The hardware and software environment is built, and the experiment shows that the recognition rate of flame recognition fusion algorithm is more than 99% under normal illumination, and has strong adaptability to complex background and high light environment.
出处
《工业技术创新》
2017年第6期15-18,共4页
Industrial Technology Innovation
关键词
变电站
火焰识别
动态特征
静态特征
Substation
Flame Recognition
Static Characteristics
Dynamic Characteristics
