基于智能图像的变电站设备绝缘子破裂裂纹识别研究  

Research on substation equipment insulator crack identification based on intelligent images

作　　者：姚楠 蔡越 苗佳[2] 

机构地区：[1]国网江苏省电力科学研究院,江苏南京211103 [2]内蒙古农业大学职业技术学院,内蒙古包头014100

出　　处：《现代电子技术》2017年第22期176-178,共3页Modern Electronics Technique

基　　金：国家自然科学基金(41161045)

摘　　要：为了提高变电站设备绝缘子的故障诊断能力,提出一种基于智能图像信息融合和边缘轮廓分割检测的变电站设备绝缘子破裂裂纹识别方法。采用激光三维扫描技术进行变电站设备绝缘子的图像采集,对采集的绝缘子破裂裂纹图像进行图像信息融合处理,采用三维区域轮廓扫描方法进行变电站设备绝缘子破裂裂纹的边缘轮廓分割检测和信息增强,在仿射不变区域内对变电站设备绝缘子破裂裂纹角点分布信息进行直方图均衡化处理,实现变电站设备绝缘子破裂裂纹的识别。仿真结果表明,采用该方法进行变电站设备绝缘子破裂裂纹识别的准确性较高,输出图像的峰值信噪比较高,说明对破裂裂纹图像识别的性能较好。In order to improve the fault diagnosis capability of substation equipment insulators,a crack identification method based on intelligent image information fusion and edge contour segmentation detection is proposed for detecting cracks on substation equipment insulators.The laser 3 D scanning technology is adopted to collect images of substation equipment insulators and perform information fusion for the collected images in which insulators with cracks are identified.The 3 D area contour scanning method is adopted to perform edge contour segmentation detection and information enhancement for cracks on substation equipment insulators.The crack corner distribution information of substation equipment insulators are processed with histogram equalization in affine invariant areas to realize the crack identification of substation equipment insulators.The simulation results show that the method has high accuracy for crack identification of substation equipment insulators and high peak signal-to-noise ratios(SNRs)of output images,which indicates its good identification performance of images with cracks on the insulators.

关 键 词：智能图像 变电站设备 绝缘子 破裂裂纹识别 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置] TP391[自动化与计算机技术—控制科学与工程]