摘要
在进行绝缘子污秽处理研究时,对绝缘子进行匹配识别后不可避免的要对绝缘子进行精确定位分割,针对运用普通的Chan-Vese(CV)模型对绝缘子图像分割时存在无法完整分割目标,甚至错误分割的问题,提出应用改进的CV模型对绝缘子图像进行精确定位。首先为抑制图像过渡部分的突变,利用中值滤波增强原始图像的平滑性;其次通过分析目标图像RGB各分量特点对目标图像进行粗定位,得到绝缘子的位置点,同时运用Hough变换确定绝缘子倾斜角度;然后提出应用绝缘子的位置点与倾斜角度对绝缘子进行精确定位,并构造绝缘子的CV模型初始轮廓,最终完成绝缘子的分割处理。实验结果表明:通过改进CV模型对任意位置的绝缘子进行精确定位的方法可准确、快速地分割绝缘子图像,具有较强的自适应性。
In the study of insulator pollution treatment,it is inevitable to accurately locate the insulator after matching and identifying the insulator.In order to solve the problem of incomplete segmentation and even segmentation while using the ordinary Chan-Vese(CV)model to segment the insulator image,an improved CV model is proposed to accurately locate the insulator in image.Firstly,in order to suppress the mutation of the transition part of the image,the median filtering is used to enhance the smoothness of the original image.Secondly,the target image is coarsely located by analyzing the characteristics of each component of the target image RGB to obtain the location of the insulator,and the tilt angle of the insulator is determined by Hough transform.Then,it is proposed to accurately location the insulator by applying the location and inclination angle of the insulator,and construct the initial contour of the CV model of the insulator,and finally complete the division processing of the insulator.The experimental results show that the method of accurately locating the insulator at any location by improving the CV model can accurately and quickly segment the insulator image,which has strong adaptability.
作者
卞建鹏
李华伟
宋海旺
郝嘉星
赵帅
孙晓云
BIAN Jianpeng;LI Huawei;SONG Haiwang;HAO Jiaxing;ZHAO Shuai;SUN Xiaoyun(University of Electrical and Electronics Engineering Shijiazhuang Tiedao University,Shijiazhuang 050043,China)
出处
《电瓷避雷器》
CAS
北大核心
2020年第4期195-201,共7页
Insulators and Surge Arresters
基金
国家自然科学基金(编号:51307112,51674169)。
作者简介
卞建鹏(1982-),男,副教授,主要研究方向:电力系统绝缘子污秽检测与识别、电气设备绝缘在线监测与故障诊断、电气设备全寿命周期管理等。
