Affiliation:
1. MOE Key Laboratory of Optoelectronic Imaging Technology and System, Beijing Institute of Technology, Beijing 100081, China
Abstract
In order to meet the fast and accurate automatic detection requirements of equipment maintenance in railway tunnels in the era of high-speed railways, as well as adapting to the high dynamic, low-illumination imaging environment formed by strong light at the tunnel exit, we propose an automatic inspection solution based on panoramic imaging and object recognition with deep learning. We installed a hyperboloid catadioptric panoramic imaging system on an inspection vehicle to obtain a large field of view as well as to shield the high dynamic phenomena at the tunnel exit, and proposed a YOLOv5-CCFE object detection model based on railway equipment recognition. The experimental results show that the mAP@0.5 value of the YOLOv5-CCFE model reaches 98.6%, and mAP@0.5:0.95 reaches 68.9%. The FPS value is 158, which can meet the automatic inspection requirements of railway tunnel equipment along the line and has high practical application value.
Funder
National Natural Science Foundation of China
Ministry of Science and Technology of China
Subject
Electrical and Electronic Engineering,Biochemistry,Instrumentation,Atomic and Molecular Physics, and Optics,Analytical Chemistry
Reference33 articles.
1. DRCDN: Learning deep residual convolutional dehazing networks;Zhang;Vis. Comput.,2020
2. Janiszewski, M., Torkan, M., Uotinen, L., and Rinne, M. (2022). Rapid Photogrammetry with a 360-Degree Camera for Tunnel Mapping. Remote Sens., 14.
3. Gavilán, M., Sánchez, F., Ramos, J.A., and Marcos, O. (2013, January 9–11). Mobile inspection system for high-resolution assessment of tunnels. Proceedings of the 6th International Conference on Structural Health Monitoring of Intelligent Infrastructure, Hong Kong, China.
4. Yasuda, T., Yamamoto, H., Enomoto, M., and Nitta, Y. (2020, January 17–18). Smart Tunnel Inspection and Assessment using Mobile Inspection Vehicle, Non-Contact Radar and AI. Proceedings of the International Symposium on Automation and Robotics in Construction, Kitakyushu, Japan.
5. Railway Tunnel Clearance Inspection Method Based on 3D Point Cloud from Mobile Laser Scanning;Yuhui;Sensors,2017
Cited by
1 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献