一种基于局部特征的PCNN电力故障区域提取方法  

Extraction Method of PCNN Electronic Equipment Fault Region Based on Local Feature

作　　者：谷凯凯[1,2] 周东国[3] 许晓路[1,2] 蔡炜[1,2] 周正钦[1,2] 胡文山 黄华[4] GU Kaikai;ZHOU Dongguo;XU Xiaolu;CAI Wei;ZHOU Zhengqin;HU Wenshan;HUANG Hua(Wuhan NARI Limited Liability Company of State Grid Electric Power Research Institute,Wuhan 430074,China;Nanjing NARI Group Corporation of State Grid Electric Power Research Institute,Nanjing 211000,China;College of Power and Mechanical Engineering,Wuhan University,Wuhan 430072,China;State Grid Shanghai Electric Power Company,Shanghai 200120,China)

机构地区：[1]国网电力科学研究院武汉南瑞有限责任公司,武汉430074 [2]国网电力科学研究院南京南瑞集团公司,南京211000 [3]武汉大学动力与机械学院,武汉430072 [4]国网上海市电力公司,上海200120

出　　处：《计算机工程》2018年第7期291-296,共6页Computer Engineering

基　　金：国家电网公司总部科技项目"基于移动应用平台的设备状态数据流录入技术研究与试点应用"(524606160034)

摘　　要：为有效提取红外图像中电气设备的故障或异常区域,提出一种新的红外图像故障区域提取方法。以脉冲耦合神经网络(PCNN)同步点火机理为依据,通过简化其内部参数,同时在参数优化配置下结合故障区域和非故障区域邻域边界的局部特征,设置PCNN模型迭代结束规则,从而使模型能进行自适应迭代,并获取红外图像故障区域。针对实际的红外检测图像进行实验,结果表明,与Otsu、k-means、分水岭及改进的PCNN方法相比,该方法具有较好的故障区域提取性能。In order to extract the fault and abnormal electronic equipment in infrared image,this paper presents a new method to extract the fault region in the infrared image. Based on the mechanism of synchronize pulse of Pulse Coupled Neural Network（ PCNN）,the inner parameters are simplified partly. At the same time,the local feature of the neighborhood of fault area and non-fault area are combined with parameter optimization configuration, and the stop rule for PCNN model iteration is set, so that the model can adaptively iterate and get the infrared image fault area. According to the actual infrared detection image,experimental results show that compared with Otsu,k-means,watershed and improved PCNN methods,the proposed method has better fault area extraction performance.

关 键 词：红外检测 脉冲耦合神经网络模型 故障区域 同步点火 局部特征 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]