With the burgeoning development of railway system throughout the world, accurate and efficient monitoring the state of rail tracks, a method serves one part of ensuring the safe operation of the railway system, are becoming increasingly important. Therefore, monitoring the health of railway track plays an indispensable role during the management of railway system. The appearance of convolutional neural network (CNN) greatly improves the problems of low accuracy and speed of traditional defect detection technology [1]. In this context, a method of rail health monitoring based on convolutional neural network is proposed in this paper. Series of You Only Look Once (YOLO) algorithms were applied to the detection of railway tracks defects, and a modified model (YOLOv3-M) based on YOLOv3 was proposed. In order to verify its effectiveness, experiments were conducted, and the results illustrated that the proposed method can effectively monitor the railway track state prior to its fail.
Railway Tracks Defects Detection Based on Deep Convolution Neural Networks
Lect. Notes Electrical Eng.
09.02.2021
11 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Detection of Defects in the Railway Tracks Based on YOLOv5
Springer Verlag | 2023
|Multi-sensor obstacle detection on railway tracks
IEEE | 2003
|