Datasets
Standard Dataset
RSDDs
- Citation Author(s):
- Submitted by:
- QIKE WU
- Last updated:
- Thu, 11/07/2024 - 01:02
- DOI:
- 10.21227/qtv6-n081
- License:
20 Views- Categories:
- Keywords:
Abstract
The Railway Surface Defect Detection (RSDDs) dataset was created to enhance the safety and reliability of railway transportation. This dataset comprises two subsets: Type-I RSDDs and Type-II RSDDs, which were collected from express and common/heavy haul railways, respectively. Type-I RSDDs consists of 67 images, each measuring 160×1000 pixels, while Type-II RSDDs includes 128 images, each measuring 55×1250 pixels. These images were captured under various lighting conditions to simulate real-world railway manufacturing and maintenance environments. Each image has been meticulously annotated by professionals, ensuring the accuracy of defect detection. The RSDDs dataset is designed to support the development and evaluation of deep learning models for the automatic identification and classification of various defects on railway surfaces, such as cracks, pores, and wear. This dataset holds significant practical value for the automation of inspection and maintenance in the railway industry.
The RSDDs dataset is a specialized collection focused on the detection of defects on railway surfaces, comprising two subsets: Type-I RSDDs and Type-II RSDDs. The Type-I RSDDs subset is sourced from express rails and consists of 67 images, each with dimensions of 160×1000 pixels. In contrast, the Type-II RSDDs subset is derived from common/heavy haul rails and includes 128 images, each measuring 55×1250 pixels. Images from both datasets were captured under diverse lighting conditions to mimic the manufacturing processes on actual assembly lines, and each image contains at least one defect, characterized by complex backgrounds and substantial noise. Defects within the dataset have been annotated by professionals in the field of rail surface detection. The purpose of the RSDDs dataset is to facilitate the training of deep learning models for the detection of defects on railway surfaces, such as cracks, pores, and wear. The high-quality images and professional annotations make it a valuable resource for research into railway surface defect detection.
