Railway Track Misalignment Detection Image Dataset

The dataset has railway track images of two types: normal and defective

The task is to classify a given image into normal or defective.

The defective images have misalignment problems: buckling or hogging

If you find this dataset useful in your research, please cite the below paper:

@INPROCEEDINGS{9643106,

author={Jerripothula, Koteswar Rao and Ansari, Sharik Ali and Nijhawan, Rahul},

booktitle={2021 34th SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI)},

title={A Vision-based Solution for Track Misalignment Detection},

year={2021},

pages={271-277},

keywords={Graphics;Annotations;Railway accidents;Transfer learning;Buildings;Feature extraction;Rail transportation;railway;transfer learning;VGG;Inception;buckling;hogging;misalignment},

doi={10.1109/SIBGRAPI54419.2021.00044}

}