基于机器视觉的绝缘子RTV涂层质量检测方法研究及应用  

Research and Application of RTV Coating Quality Inspection Method Based on Machine Vision

作　　者：申振腾 陈建平[1] 杜厚栋 李瑞 SHEN Zhenteng;CHEN Jianping;DU Houdong;LI Rui(School of Mechanical Engineering,Tianjin University of Science ＆ Technology,Tinjin 300222,China)

机构地区：[1]天津科技大学机械工程学院,天津300222

出　　处：《电瓷避雷器》2018年第5期233-240,共8页Insulators and Surge Arresters

基　　金：国家电网天津市东丽供电分公司合作项目(编号:KJ15-1-16)

摘　　要：输变电路绝缘子室温硫化硅橡胶涂料(room temperature vulcanized silicone rubber coating,RTV)涂层质量缺陷严重影响着绝缘子的工作状态,但是目前尚无有效的方法对其在运行过程中可能出现的质量问题进行检测。为此提出一种基于机器视觉技术的绝缘子RTV涂层质量检测方法。该方法通过色彩空间转换、加权滤波、阈值分割、边缘检测及形态学运算等一系列算法,实现了理论上RTV涂层质量检测。利用Halcon、QT、C++等混合编程方法,设计研制一套检测装置,在单个和成串情况下,对该方法的检测效率、准确性等评价因素进行了验证。实验证明该方法缺陷检测方便快捷、准确性高,可用于工程现场检测。RTV coating quality defects seriously affects the working state of the insulator,but there is no effective method to detect the quality problems that may arise in operating. Due to the deficiency,a method for quality inspection of insulator RTV coating relaying on machine vision technology was proposed. The method realizes the detection of insulator RTV coating quality by a series of algorithms such as color space conversion,weighted filtering,threshold segmentation,edge detection and morphological operation and so on. Thena detection system,in a single and string of cases,was designed and developed to varify the detection efficiency,accuracy and other factors of evaluation on the approach by using Halcon,QT,C ＋＋ and others mixed programming method. Experiments show that the defect detection means is convenient and high accuracy,and can be used to project site testing.

关 键 词：RTV涂层质量 机器视觉 图像分割 边缘检测 形态学 

分 类 号：TM216[一般工业技术—材料科学与工程]