RSDDs dataset

The goal to create the RSDDs dataset is to provide the worldwide researchers of rail inspection community a database for evaluating their algorithms. We anticipate this will serve as standardized datasets for comparing techniques, something which has greatly helped other defects detection algorithms but has largely been lacking in rail surface inspection field.

RSDDs dataset contains two types of dataset: the first is Type-I RSDDs dataset captured from express rails, which has 67 challenging images. The second is Type-II RSDDs dataset captured from common/heavy haul rails, which has 128 challenging images. Note that each image from these two datasets contains at least one defect, and have complex background with much noise. These defects in the RSDDs dataset are marked by some professional human observers in the rail surface inspection field, and stored in the folds. Some typical images are displayed as below.

Type-I RSDDs dataset

Type-II RSDDs dataset

Download notice:
Download the database and corresponding groundtruth ( rar file ), and please send an email to Qingyong Li (

If you use this database, please cite the following publication:

Jinrui Gan, Q.Y. Li , et al. A Hierarchical Extractor-Based Visual Rail Surface Inspection System. IEEE Sensors Journal, 2017, 17(23):7935-7944.