基于改进Canny的弓网燃弧识别算法研究  

Study on Pantograph-contact Arcing Identification Algorithm Based on Improved Canny Method

作　　者：龚石琅 苏燕辰[1] 杨岗[1] GONG Shilang;SU Yanchen;YANG Gang(School of Mechanical Engineering,Southwest Jiaotong University,Chengdu 610031,China)

机构地区：[1]西南交通大学机械工程学院,四川成都610031

出　　处：《铁道车辆》2024年第6期192-197,共6页Rolling Stock

基　　金：国家重点研发计划(2020YFB1200300ZL);四川省重点研发项目(2023YFG0063)。

摘　　要：针对使用传统Canny算法进行弓网燃弧检测导致的检测率低的问题,文章以地铁燃弧图像为试验目标,提出了基于改进Canny算法的弓网燃弧检测技术。针对传统Canny算法中高斯滤波导致图片模糊的问题,提出了双边均值迁移滤波(Bilateral Mean Shift Filtering,BMSF)进行滤波处理取代传统的高斯滤波;针对手动输入阈值导致算法适应性差的问题,引入了Otsu算法改进传统的双阈值算法;针对部分燃弧过小而引起的漏检现象,将处理后的二值图像进行膨胀操作,增强燃弧的边界信息。将改进算法与传统算法进行仿真试验评估。结果表明在普通环境与强光环境下,准确率都有一定程度的提升,为基于图像处理的燃弧检测方面的研究提供了新思路。To solve the problem of low detection rate caused by applying the conventional Canny method to the detection of pantograph-contact arcing,the article takes the metro arcing image for the object of test and proposes the pantograph-contact arcing detection technology based on the improved Canny method.For the problem of blurred pictures caused by the Gaussian filter in the conventional Canny method,it proposes to replace the conventional Gussian filter with the bilateral mean shift filtering(BMSF)for filter processing;for the problem of bad adaptability of algorithm caused by manual input of threshold,it introduces the Otsu algorithm to improve the conventional dual threshold method;for the omission of detection caused by some undersize arcing,it expands the processed binary image to strengthen the boundary information of arcing.The improved algorithm and the conventional algorithm are evaluation by a simulation test.The result shows that the accuracy rate is improved to a certain extent in both ordinary environment and strong light,providing a new idea for the study on arcing detection based on image processing.

关 键 词：电力系统 CANNY算法 弓网燃弧 双边均值滤波 双阈值算法 

分 类 号：U270[机械工程—车辆工程]