摘要
针对变电站带电水冲洗机器人作业时的绝缘子串自主识别问题,基于机器人视觉系统获得的光学图像,研究提取出图像的灰度特征曲线,并分析了图像的灰度周期性特征和灰度阈值特征;采用D-S证据理论,提出了一种基于图像灰度特征融合的变电站绝缘子串自主识别方法;利用带电水冲洗机器人机器视觉系统单元测试平台对所提出的方法进行了现场实验验证,实验结果表明该方法对室外光照条件和复杂背景干扰具有良好的适应性,能够实现对单段、双段两类支柱缘子串的准确识别。
Aiming at the problem of self-identification of insulator strings in substation charged water washing robot operation,based on the optical image obtained by robot vision system,the gray characteristic curve of the image is extracted in the research,and the gray periodic features and gray threshold characteristics of the image are analyzed.Based on D-S evidence theory,a self-identification method for substation insulator string based on image gray feature fusion is proposed.The proposed method was validated by field test using the charged machine water washing robot vision system unit test platform.The experimental results show that the method has good adaptability to outdoor lighting conditions and complex background interference,and it also can realize the accurate identification of single segment and double segment pillar margin substrings.
作者
李明华
姜勇
顾光荣
岳湘
LI Ming-hua;JIANG Yong;GU Guang-rong;YUE Xiang(Guizhou Power Grid co.,LTD.Kaili Power Supply Bureau,Kaili 556000,China;State Key Laboratory of Robotics,Shenyang Institute of Automation,Chinese Academy of Sciences,Shenyang 110016,China)
出处
《电子设计工程》
2019年第23期1-6,共6页
Electronic Design Engineering
基金
国家自然科学基金重点项目(5153000289)
关键词
带电水冲洗机器人
绝缘子串
特征提取
D-S证据理论
自主识别
charged water washing robot
insulator string
feature extraction
D-S evidence theory
self-identification
