摘要
传统的火灾检测系统由于传感器受环境影响较大导致效果较差。针对这一问题,文中基于图像处理技术,结合红外基础理论、双目立体视觉测距理论和开源计算机视觉库(OpenCV)设计并实现了一套火灾火源定位及检测系统。该系统能根据红外测温、轴向比判别、内外焰判别、形状不规则判别等算法,对火灾燃烧时出现的典型图像特征进行检测和报警,并能仿照双目视觉定位原理对火源进行定位与跟踪。该系统经过实验测试,检测和定位效果良好,能为类似的火灾检测系统的设计提供技术参考。
Traditional fire detection system is easily affected by the environment,so combined with the infrared basic theory,binocular vision theory as well as the open source computer vision library(OpenCV),a new fire source location and detection system based on image processing technology is designed and implemented by this paper.The system can detect fire by recognizing the typical image features that occur when the fire is burned according to algorithms of the infrared temperature measurement,the axial ratio discrimination,the internal and external flame discrimination and the irregular discrimination,then alert administrators in time.The fire source can also be positioned and tracked by this system,following the binocularvisiontheory.Thissystemistested,operating well in detecting and positioning,providing technical reference for the design of the similar fire detection system.
作者
储海东
赵岩
庄斌
王艳如
杨璇
CHU Hai dong;ZHAO Yan;ZHUANG Bin;WANG Yan ru;YANG Xuan(State Grid Corporation of China,Beijing 100031,China;Beijing Guodiantong Network Technology Co.,Ltd.,Beijing 100070,China)
出处
《电子设计工程》
2020年第7期156-160,164,共6页
Electronic Design Engineering
基金
国家电网公司总部科技项目(52110118008H)。
