基于快速鲁棒性特征匹配的电气化铁路绝缘子不良状态检测.pdfVIP

第37 卷 第8 期 电 网 技 术 Vol. 37 No. 8 2013 年8 月 Power System Technology Aug. 2013 文章编号：1000-3673 （2013 ）08-2297-06 中图分类号：TM 835.4 文献标志码：A 学科代码：470·4034 基于快速鲁棒性特征匹配的 电气化铁路绝缘子不良状态检测 杨红梅，刘志刚，韩烨，韩志伟 （西南交通大学电气工程学院，四川省 成都市 610031 ） Defective Condition Detection of Insulators in Electrified Railway Based on Feature Matching of Speeded-up Robust Features YANG Hongmei, LIU Zhigang, HAN Ye, HAN Zhiwei (School of Electrical Engineering, Southwest Jiaotong University, Chengdu 610031, Sichuan Province, China) ABSTRACT: In allusion to automatic identification and 以人工方式定期对绝缘子进行清扫和检查，这种方 location of defective insulators in overhead contact line of 式很难实时监控绝缘子状态。在电力系统中，基于 electrified railway by non-contact image detection, a detection 图像处理技术的绝缘子检测方法的研究相对较多， method based on feature matching of speeded-up robust features (SURF) is proposed. Firstly, the feature extraction and 如：何洪英等结合图像处理和人工智能技术实现了 matching of insulator in the image to be detected are performed 绝缘子污秽特征提取及污秽等级检测[1-2] ；李佐胜等 by SURF; secondly, the pre-processing of insulator such as 研究了基于方差分析的绝缘子特征选取方法[3] ；黄 angle correction and morphological operation is carried out; 宵宁等通过最大熵阈值方法进行图像分割，应用连 finally, longitudinal statistics of insulator grayscale is 通区域方法对绝缘子串进行识别与提取[4] ；林聚财 implemented and according to the regularity of distribution of 等利用颜色和直方图匹配方法识别玻璃绝缘子损 the minimal value of the grayscale the defective insulator can be identified and located. Experimental results show that using 伤区域[5] ；姚建刚等利用相对温度分布特征和人工 the proposed method the defective insulators can be identified 神经