摘要
软质绝缘遮蔽用具的表面缺陷对其绝缘性能和使用寿命有重要影响,为解决人工检测漏检、误检率高的问题,文中提出基于机器视觉的绝缘毯表面微缺陷智能无损检测方法。针对不同环境下图像特点,选取改进Sauvola算法对图像进行局部阈值分割,结合形态学操作提取特征,实现表面微缺陷的识别;通过搭建沿面放电试验平台获取不同微缺陷对其沿面放电的影响,采用皮尔逊相关性系数分析了微缺陷的单个最大缺陷面积比和全部缺陷面积比与绝缘毯表面放电电压之间的相关性。结果表明:文中方法识别准确率95.3%,单次检测时间0.53 s,提升了检测准确性与效率;单个缺陷面积的大小与闪络电压呈负相关且为极强相关,单个缺陷越大沿面闪络电压下降越多,沿面闪络电压平均下降5%~22%,验证了识别结果的有效性,可为绝缘毯的现场外观检测提供依据。
The surface defects of EVA insulating cover have an important influence on the insulation performance and service life.In order to solve the problem of high error detection rate and missing detection rate in manual detec⁃tion,the detection method of insulating blanket surface micro⁃defects based on machine vision is proposed in this pa⁃per.For the characteristics of images in different environments,the improved Sauvola algorithm is selected to per⁃form local threshold segmentation on images and,in combination with morphological operation,operation feature of extraction to achieve the identification of surface micro⁃defects.The influence of different micro⁃defects on the sur⁃face discharge of insulating blanket is obtained by setting up the surface discharge test platform.The correlation be⁃tween the single maximum defect area ratio and the total defect area ratio and the surface discharge voltage of insulat⁃ing blanket is analyzed by using Pearson correlation coefficient.The results show that the identification accuracy rate of the method proposed in this paper is 95.3%and the single detection time is 0.53 s,which improve the detection ac⁃curacy and efficiency.The size of single defect area is negatively correlated with the flashover voltage and it is strong⁃ly correlated.The larger the single defect is,the more the flashover voltage decreases,and the flashover voltage along the surface decreases by 5%~22%on average,which verifies the effectiveness of the identification results and pro⁃vides a base for the on⁃site appearance inspection of insulation blanket.
作者
吴田
罗成军
王申华
杨运国
王旭杰
WU Tian;LUO Chengjun;WANG Shenhua;YANG Yunguo;WANG Xujie(Hubei Provincial Engineering Technology Research Center for Power Transmission Line,China Three Gorges University,Hubei Yichang 443002,China;College of Electrical Engineering&New Energy,China Three Gorges University,Hubei Yichang 443002,China;State Grid Zhejiang Wuyi County Power Supply Co.,Ltd.,Zhejiang Jinhua 321000,China)
出处
《高压电器》
CAS
CSCD
北大核心
2022年第11期75-81,共7页
High Voltage Apparatus
基金
国家自然科学基金(51807110)。
关键词
绝缘毯
表面缺陷检测
机器视觉
Sauvola算法
沿面放电
insulation blanket
surface defects detection
machine vision
Sauvola algorithm
surface discharge
