摘要
在电气化铁路接触网设备中,绝缘子是最易发生故障的零部件之一。文章运用接触网成像检测图片,采用SVM分类器,基于HOG特征对接触网成像数据中的绝缘子进行初步定位,再使用局部二值化实现绝缘子精确定位。针对绝缘子破损缺片、闪络放电、附着异物等3种典型缺陷,分别设计了故障识别算法,且该方法已在中国铁路广州局集团有限公司部分线路中进行实测,效果良好。
Insulator is one of the most vulnerable parts in OCS of electrified railway. This paper uses the image of OCS imaging detection and SVM classifier, and the insulators in the OCS imaging data are positioned preliminarily based on the HOG features and then the precise location of insulators is realized by local binarization. Aiming at three typical defects of insulator, such as broken insulator, flashover discharge and foreign object attachment, fault identification algorithms are designed respectively. The method has been tested on some lines of Guangzhou Railway Administration Group Co., Ltd. China Railway, and the results are satisfactory.
出处
《现代城市轨道交通》
2019年第10期5-10,共6页
Modern Urban Transit
