基于图像识别的高压放电紫外检测研究  

Study on Ultraviolet Detection of High Voltage Discharge Based on Image Recognition

作　　者：路永玲 王沛渊 胡成博 陶风波 王真 LU Yonglin;WANG Peiyuan;HU Chengbo;TAO Fengbo;WANG Zhen(State Grid Jiangsu Electric Power Co,Ltd.Electric Power Science Research Institute(State Grid Electric Power System Artificial Intelligence Joint Laboratory),Nanjing 211103,China;Tsinghua Shenzhen International Graduate School,Shenzhen 518000,China)

机构地区：[1]国网江苏省电力有限公司电力科学研究院(国家电网电力系统人工智能联合实验室),南京211103 [2]清华大学深圳国际研究生院,广东深圳518000

出　　处：《自动化与仪器仪表》2023年第2期66-69,75,共5页Automation & Instrumentation

基　　金：国网江苏省电力有限公司2021年科技项目《输电通道可视化及物联网边缘计算融合终端研制》(JF2021031)。

摘　　要：紫外检测技术可以实现高压电气设备电晕放电带电检测,便于检测人员及时了解设备的绝缘性能。利用模糊C均值聚类算法和数学形态学滤波将紫外检测图像中的放电区域提取和滤波,并计算各区域面积。通过试验对影响放电区域面积的检测距离和增益因素进行了研究,对检测距离和增益因素进行了校正,为高压电气设备电晕放电评估系统提供了一定的指导。Ultraviolet detection technology can realize corona discharge charged detection of electrical equipment, which is convenient for testing personnel to know the insulation performance of equipment in time. Fuzzy C-means clustering algorithm and mathematical morphological filtering were used to extract and filter the discharge region in the UV detection image, and the area of each region was calculated. The detection distance and gain factors that affect the discharge area are studied through experiments, and the detection distance and gain factors are corrected, which provides a certain guiding significance for corona discharge evaluation of electrical equipment.

关 键 词：紫外检测 图像处理 滤波 光子数 检测距离 增益 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]