基于改进C-V模型的外绝缘放电紫外图像特征量提取  

Ultraviolet Image Parameters Extraction of Insulation Surface Discharge Based on Improved C-V Model

作　　者：王丰华[1] 刘国坚 张宏钊 黄荣辉 刘亚东[1] WANG Fenghua;LIU Guojian;ZHANG Hongzhao;HUANG Ronghui;LIU Yadong(Key Laboratory of Control of Power Transmission and Conversion（Shanghai Jiao Tong University）,Ministry of Education,Shanghai 200240,China;College of Electric Engineering,Shanghai University of Electric Power,Shanghai 20090,China;Shenzhen Power Supply Bureau Co.,Ltd.,Shenzhen 518000,China)

机构地区：[1]电力传输与功率变换控制教育部重点实验室(上海交通大学),上海200240 [2]上海电力学院电气工程学院,上海200090 [3]深圳供电局有限公司,深圳518000

出　　处：《高电压技术》2018年第8期2525-2532,共8页High Voltage Engineering

基　　金：国家自然科学基金(51307109)~~

摘　　要：为实现紫外（UV）成像技术检测绝缘子缺陷时产生的紫外图像中放电区域的自动准确分割,提出了一种基于改进Chan-Vese（C-V）模型的放电光斑提取方法。通过引入符号距离保持项、用水平集函数梯度的模替代Dirac函数以及简化模型参数来对现有C-V模型进行了改进,然后将其用于获取绝缘子表面放电紫外图像的放电光斑,并给出了基于常规边缘检测算子法、数学形态学的紫外图像分割结果。结果表明：改进后的C-V模型能更为快速准确地分割出紫外成像的目标光斑,且具有较好的抗干扰能力。与现有方法相比较,使用改进的C-V模型来提取放电光斑的准确度提高了5%,且对C-V模型中关键参数的依赖性较小,非常适合于紫外图像特征量的提取。研究结果可为输电线路外绝缘的有效检测提供重要依据。To segment the discharge area of insulator in ultraviolet（UV） image produced by ultraviolet imaging technology automatically and accurately, we proposed an improved method based on an improved Chan-Vese（C-V） model to analyze the discharge area, where the existing C-V model is improved by adding a penalized energy term, displacing the Dirac function with the norm of level set function gradient, and simplifying amount of parameters. Then, the method was used to obtain the discharge spot of the ultraviolet image of the insulator surface, and the UV image segmentation results based on the conventional edge detection algorithm and mathematical morphology were also given. The research results indicate that the proposed method can segment the discharge area faster and more accurately with stronger anti-interference ability. The accuracy of the improved C-V model is 5% higher than that of current methods, and the dependency on key parameters is lower. The proposed C-V model is excellent for segmentation of ultraviolet image parameters. The research results can provide an important reference for the effective detection of the insulation for the transmission line.

关 键 词：边缘检测 C-V模型 放电区域 光斑面积 图像分割 外绝缘 紫外成像 

分 类 号：TM855[电气工程—高电压与绝缘技术] TP391.41[自动化与计算机技术—计算机应用技术]