Machine Vision-based Research on the Inspection of Dropper Defects of Overhead Contact Line
LIU Jie1, XU Jianguo1, GAO Chunli2, LIU Qiuhang3
1. China Railway Electrification Bureau Group Co. Ltd, Beijing 100036, China; 2. Dalian Weide Integrated Circuit Co. Ltd, Dalian, Liaoning 116085, China; 3. Beijing Institute of Technology, Beijing 100081, China
Abstract：Research purposes: To solve the problem of low efficiency and long time due to manual inspection, we create an intelligent inspection system of dropper defects of overhead contact line by applying high-definition image acquisition technology and image defect recognition modelling technology. Research conclusions: (1) By introducing visual intelligence technology, the research can establish the image detection standard for the installation status of overhead contact line dropper, create a database of defect images of overhead contact line dropper, and perform dynamic intelligent and rapid recognition and detection of overhead contact line dropper defect status. (2) The system uses the Faster R-CNN neural network model to train the defect data of overhead contact line dropper. With data transformation processing on the image data, the mean average precision (mAP) of the training results can reach 81.36%. (3) The system has high recognition rate accuracy, whichcan improve the detection efficiency and reduce omission. (4) The research results can provide a new way to change the traditional overhead contact line inspection method into intelligent inspection, and have certain reference value.
刘杰, 许建国, 高春丽, 刘裘航. 基于机器视觉的接触网吊弦缺陷检测研究[J]. 铁道工程学报, 2022, 39(5): 91-97.
LIU Jie, XU Jianguo, GAO Chunli, LIU Qiuhang. Machine Vision-based Research on the Inspection of Dropper Defects of Overhead Contact Line. Journal of Railway Engineering Society, 2022, 39(5): 91-97.
赵晓娜,吴兴军,徐根厚.德国高速铁路接触网检测系统[J].中国铁路,2008(9):60-62.Zhao Xiaona, Wu Xingjun, Xu Genhou. Detection System of Overhead Contact Line of German High-speed Railway [J]. Chinese Railways, 2008(9):60-62.
[2]
Szegedy C, Toshev A, Erhan D. Deep Neural Networks for Object Detection[C]//Advances in Neural Information Processing Systems,2013: 2553-2561.
[3]
Ren S, He K, Girshick R, etc. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2017(6):1137-1149.
[4]
余晓宁,顾桂梅,王阳萍,等.基于Faster R-CNN的接触网吊弦故障检测方法[J].兰州交通大学学报,2021(2):58-65.Yu Xiaoning, Gu Guimei, Wang Yangping, etc. Catenary Dropper Fault Detection Method Based on Faster R-CNN[J]. Journal of Lanzhou Jiaotong University, 2021(2): 58-65.
[5]
李兵祖,宋超,武莹,等.基于多尺度深度学习的接触网吊弦异常检测及应用[J].电气化铁道,2020(4):42-45.Li Bingzu,Song Chao,Wu Ying, etc.Detection of Abnormality of OHL Droppers Based on Multi-scale Deep Learning and Its Application [J]. Electric Railway, 2020(4):42-45.
2022, Vol. 39 
版权所有 © 2014 《铁道工程学报》编辑部