摘要
在变电站中,当变压器遭遇火灾时,火焰与烟雾区域的边缘往往呈现出细微的纹理与形态变化。以上变化与火焰和烟雾区域的相互交织,共同导致了火灾源头的精确位置难以被识别,对应急救援工作的目标定位造成了阻碍。为此,提取细微纹理特征,提出变电站变压器火灾点定位方法。结合RGB空间原理,分割变电站变压器的火灾区域与烟雾区域,计算每个通道的颜色矩,引入LBP算法获取相邻像素点之间的灰度差异,提取烟雾细微纹理特征,构建特征集合,并将其作为支持向量机的输入,准确定位变电站变压器火灾点。通过仿真结果可知,所提方法成功定位出了火灾发生点,交并比平均值高达0.902,由此说明上述方法在火灾点应急定位任务中的有效性。
In the substation,when the transformer encounters a fire,the edges of the flame and smoke areas often show subtle texture and morphological changes.The above changes are intertwined with the flame and smoke areas,which together make it difficult to identify the exact location of the fire source,and hinder the target positioning of emergency rescue work.For this reason,fine texture features are extracted,and a fire location method for transformers in substations is proposed.Combined with the principle of RGB space,the fire area and smoke area of the substation transformer are divided,the color moment of each channel is calculated,LBP algorithm is introduced to obtain the gray difference between adjacent pixel points,extract the fine texture features of smoke,build a feature set,and use it as the input of support vector machine to accurately locate the fire point of substation transformer.The simulation results show that the proposed method successfully locates the fire occurrence point,and the average value of the intersection and merger ratio is as high as 0.902,which shows the effectiveness of the above method in the emergency location task of the fire point.
作者
李夏川
路朝阳
李孟
陈晓童
LI Xia-chuan;LU Zhao-yang;LI Meng;CHEN Xiao-tong(Qingdao Power Supply Company of State Grid Shandong Electric Power Company,Qingdao Shandong 266101,China;School of Mechanical and Electrical Engineering,China University of Petroleum,Qingdao Shandong 266580,China)
出处
《计算机仿真》
2025年第4期78-82,共5页
Computer Simulation
基金
国家电网公司青岛供电公司科学技术项目(SGTYHT/20-JS-223)。
关键词
大型变电站
变压器
火灾点
区域分割
特征提取
应急定位方法
Large substation
Transformer
Fire point
Area segmentation
Feature extraction
Emergency localization method
