特征融合的电力机械设备过热故障红外检测  

Infrared Detection of Overheat Fault of Electric Machinery Equipment Based on Feature Fusion

作　　者：陈秋菊 彭天昊 康万杰 何国锋 CHEN Qiu-ju;PENG Tian-hao;KANG Wan-jie;HE Guo-feng(Department of Brewing Engineering Automation on Moutai Institute,Guizhou Renhuai 564507,China;School of Computer Science on Guizhou University,Guizhou Guiyang 550025,China)

机构地区：[1]茅台学院酿酒工程自动化系,贵州仁怀564507 [2]贵州大学计算机学院,贵州贵阳550025

出　　处：《机械设计与制造》2024年第4期337-341,共5页Machinery Design & Manufacture

基　　金：遵义市科技局、茅台学院联合科技研发资金项目—信息互联驱动的酿造设备故障诊断方法研究(遵市科合HZ字[2020]304号)。

摘　　要：因电力资源需求急剧上升,电力系统提高电能资源运输量的同时,电力机械设备经常出现过热问题,易导致电力故障的发生,为此提出特征融合的电力机械设备过热故障红外检测方法。将电力机械设备红外图像转换到HIS颜色空间,采用中值滤波法去噪处理红外图像,降低噪声对故障检测效果的影响。提取红外图像的反射与入射分量,获取图像相关细节信息,在信息融合的帮助下合并红外图像信息,完成红外图像增强。使用梯度方法定位图像边缘,通过划分红外图像中高温区域,实现电力机械设备过热故障检测。实验结果表明,所提方法的过热故障红外检测误差低、检测时间短,鲁棒性高。Due to the sharp rise of power resource demand,while the power system improves the transportation of power resources,the overheating problem often occurs in electrical machinery equipment,which is easy to lead to the occurrence of power failures.Therefore,an infrared detection method for overheating failures of electrical machinery equipment based on feature fusion is pro-posed.The infrared image of power machinery equipment is converted to HIS color space,and the median filtering method is used to denoise the infrared image,so as to reduce the impact of noise on the fault detection effect.Extract the reflection and incidence components of the infrared image,obtain the relevant details of the image,and combine the infrared image information with the help of information fusion to complete the infrared image enhancement.The gradient method is used to locate the edge of the im-age,and the overheating fault detection of electrical machinery equipment is realized by dividing the high temperature region in the infrared image.The experimental results show that the proposed method has low infrared detection error,short detection time and high robustness.

关 键 词：特征融合 电力机械设备 过热故障 红外检测 HIS颜色空间 梯度方法 

分 类 号：TH16[机械工程—机械制造及自动化] TP242[自动化与计算机技术—检测技术与自动化装置]