摘要
为提升火灾防治方法,通过最优阈值预处理火灾图像,利用团块标记算法分离火灾区域与背景区域,通过OpenCV库训练识别火灾区域的色度,根据火焰颜色特征复核校验目标区域,分析色度匹配误差、匹配成功率以及火灾识别率,实现了对火灾的快速识别与响应。结果表明:该算法使火灾响应时间相较于传统算法缩短了83.6%,平均时长在0.5 s左右;火灾识别准确率提升了15.7%,达到95.6%。
This paper is aimed at improving fire prevention.The study involves preprocessing fire image with optimal threshold,separating the fire area and background with the block marking algorithm,training and distinguishing the color characteristics of the fire area with OpenCV library,checking and verifying the fire area based on the characteristics of flame color,analying the chroma matching error,the matching success rate and fire recognition rate,and having the quick recognition and response realized to fire.??The results show that the fire response time is reduced by 83.6%compared with the traditional algorithm,and the average time is about 0.5s;and fire identification accuracy is increased by 15.7%,reaching 95.6%by the new algorithm.
作者
杨婷
赵杨辉
Yang Ting;Zhao Yanghui(Inner Mongolia Power Group Co.Ltd., Alxa Electirc Power Supply Company, Alxa Inner Mongolia 750306, China)
出处
《黑龙江科技大学学报》
2022年第2期263-268,共6页
Journal of Heilongjiang University of Science And Technology
关键词
火灾识别
团块标记
最优阈值
颜色特征
five recognition
block marking
optimal threshold
color characteristics
