Automatic Detection of Rail Defects from Images

Emil Hovad*, Helena Hansen, André Filipe da Silva Rodrigues, Vedrana Andersen Dahl

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingBook chapterResearchpeer-review

Abstract

In this study, images of the rails captured by a track recording car are investigated. The images are processed in order to asses and predict the rail quality by a simple method, which gives the possibility of fast and easy implementation. The first step of the method is detecting the rails from the images with an algorithm. The second step of the method is finding visually noticeable defects on the detected rail with automatized algorithms for defect detection. One of the algorithms for defect detection shows promising results and investigating the rail images could be a promising complementary method to the already used manual ultrasound measurements.

Original languageEnglish
Title of host publicationIntelligent Quality Assessment of Railway Switches and Crossing
PublisherSpringer
Publication date2021
Pages187-205
ISBN (Print)978-3-030-62471-2
DOIs
Publication statusPublished - 2021
SeriesSpringer Series in Reliability Engineering
ISSN1614-7839

Bibliographical note

Publisher Copyright:
© 2021, Springer Nature Switzerland AG.

Fingerprint

Dive into the research topics of 'Automatic Detection of Rail Defects from Images'. Together they form a unique fingerprint.

Cite this