Rail flaw detection is an essential link in the normal operation of the railway. Accurately detecting the internal damage of the rail and repairing the rail in time can find and eliminate the hidden dangers before the accident, and provide a strong security guarantee for the running of the train. In this paper, a rail defect detection method based on improved XGBoost is proposed. The Conditional Generative Adversarial Networks is used to expand the existing rail damage detection data set, then an improved XGBoost model is used to identify and classify the rail flaw detection data. Taking the crack damage type of screw hole as an example, good experimental results are obtained, which further proves the effectiveness and reliability of the scheme.
CITATION STYLE
Zhang, C., Zhao, Q., Shen, T., & Sun, B. (2022). Rail Defect Detection Method Based on Improved XGBoost. In Lecture Notes in Electrical Engineering (Vol. 961 LNEE, pp. 911–920). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-19-6901-0_94
Mendeley helps you to discover research relevant for your work.