电气工程学报 ›› 2021, Vol. 16 ›› Issue (3): 70-76.doi: 10.11985/2021.03.010

• 电力系统 • 上一篇    下一篇

扫码分享

基于对比度受限直方图均衡化的夜间接触网支柱号牌识别方法研究*

刘家军(), 钟鸣睿(), 秦梓轩()   

  1. 西安理工大学电气工程学院 西安 710048
  • 收稿日期:2021-05-10 修回日期:2021-07-30 出版日期:2021-09-25 发布日期:2021-10-29
  • 作者简介:刘家军,男,1967年生,教授,博士研究生导师。主要研究方向为电力系统监控与调度自动化、智能电网技术。E-mail: liujiajun8@126.com
    钟鸣睿,男,1997年生,硕士研究生。主要研究方向为电力系统监控与调度自动化。E-mail: 1251646762@qq.com
    秦梓轩,男,1997年生,硕士研究生。主要研究方向为电力系统监控与调度自动化。E-mail: Qinzixuan2020@163.com
  • 基金资助:
    * 陕西省教育厅服务地方专项计划(16JF020);西安市科技计划(2020KJRC0078);西安理工大学科技成果转化(104-256121815)

Research on Identification Method of OCS Pillar Number Plate at Night Based on Contrast Limited Histogram Equalization

LIU Jiajun(), ZHONG Mingrui(), QIN Zixuan()   

  1. School of Electrical Engineering, Xi’an University of Technology, Xi’an 710048
  • Received:2021-05-10 Revised:2021-07-30 Online:2021-09-25 Published:2021-10-29

摘要:

在接触网停电检修作业挂接地线中,挂接地线位置由作业票上的支柱号确定,因此,研究接触网支柱号牌的识别对挂接地线的自动化监测具有重要意义。为此,提出一种基于对比度受限自适应直方图均衡化(Contrast-limited adaptive histogram equalization,CLAHE)的支柱号识别方法。利用对比度受限自适应直方图均衡化和中值滤波预处理图像,用最小区域法和形态学变换提取号牌区域,再用连通域和模板匹配识别号牌上的支柱号字符,并在识别过程中考虑支柱号的分布规律,验证支柱号的识别结果。利用Halcon和C#联合开发的识别软件平台进行试验,对比CLAHE和直方图均衡化(Histogram equalization,HE)的均方差和峰值信噪比,结果表明这种方法大大提高了支柱号识别的准确率。

关键词: 挂接地线, 接触网支柱号牌识别, 对比度受限自适应直方图均衡化, 形态学变换, 连通域法

Abstract:

In the catenary power-off maintenance operation, the position of the hanging ground wire is determined by the pillar number on the work ticket. Therefore, the research on the identification of the pillar number plate of the catenary is of great significance for the automatic monitoring of the hanging ground wire. Therefore, a pillar number recognition method based on contrast limited adaptive histogram equalization (CLAHE) is proposed. The contrast constrained adaptive histogram equalization and median filter are used to preprocess the image, the minimum region method and morphological transformation are used to extract the number plate region, and then the connected domain and template matching are used to identify the pillar number characters on the number plate. In the recognition process, the distribution law of pillar numbers is considered to verify the recognition results of pillar numbers. Experiments are carried out on the recognition software platform jointly developed by Halcon and C#, and the mean square deviation and peak signal-to-noise ratio of CLAHE and histogram equalization are compared. The results show that this method greatly improves the accuracy of pillar number recognition.

Key words: Hanging ground wire, identification of catenary pillar number plate, contrast limited adaptive histogram equalization, morphological transformation, connected domain method

中图分类号: