Increased operating hours of railroads will lead to irreversible damage to railroad tracks, and these defects may seriously affect the safety of train operations. Due to the uncertainty and diversity of the defects of the railroad tracks, how to identify the defects quickly and effectively is very important for the safe operation of the railroad system. Traditional CNN-based detection methods require a large dataset with long training process and complex processing of image data. Eddy current is an effective method for detecting defects on railroad surfaces and a one-dimensional signal can be obtained directly from the eddy current sensor. One-dimensional signals are prevalent in industrial applications used for monitoring and control. In this paper, a method detecting railroad metal surface defect based on MobileViTv2 and eddy current is proposed. In the experiments, the railroad track anomalies are categorized into three types: Weld, Squat, and Joint. The results show that the classification efficiency of the method can reach 99%.
Rail Surface Defect Detection Based on MobileViTv2 and Eddy Current
Lect. Notes Electrical Eng.
International Conference on Electrical and Information Technologies for Rail Transportation ; 2023 ; Beijing, China October 19, 2023 - October 21, 2023
Proceedings of the 6th International Conference on Electrical Engineering and Information Technologies for Rail Transportation (EITRT) 2023 ; Kapitel : 62 ; 636-643
2024-01-09
8 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Steel rail surface crack flaw detection trolley based on eddy current principle
Europäisches Patentamt | 2022
|