RFDD: An Extended Road Facility Defect Detection Dataset

Important Notice: This dataset is directly related to the paper "Enhanced Road Infrastructure Defect Detection via Attention-Based Visual Inspection" submitted to The Journal of Real-Time Image Processing.If you use the dataset in your research, please be sure to cite the relevant papers. (Please cite both the original RDD2022 and our paper when using this dataset.) This is the RFDD (Road Facility Defect Detection) dataset, an extended version of RDD2022 for comprehensive road infrastructure inspection. The dataset includes: Key Extensions:1. Enhanced Annotations: Added "crack_on_line" annotations to original RDD2022 images2. New Defect Classes: Introduced "damaged guardrail" and "damaged traffic sign" categories3. Total 7 Classes: Comprehensive coverage of common road defects Dataset Specifications:- Total Images: 26014- Annotation Format: YOLO (txt files)- License: CC BY-SA 4.0- Based on: RDD2022 dataset (CC BY-SA 4.0) Associated Resources:- Code Repository: https://github.com/zdpf122/ECEM-YOLO-RoadDefect- Related Paper: "Enhanced Road Infrastructure Defect Detection via Attention-Based Visual Inspection" (Submitted to The Visual Computer) Usage:Designed for training and evaluating computer vision models for road defect detection, particularly useful for edge computing applications.data/├── images/│ ├── train/ # Training images│ ├── val/ # Validation images│ └── test/ # Test images└── labels/ ├── train/ # Training annotations (YOLO txt format) ├── val/ # Validation annotations └── test/ # Test annotationsCitation Requirements @article{arya2024rdd2022, title={RDD2022: A multi-national image dataset for automatic road damage detection}, author={Arya, Deeksha and Maeda, Hiroya and Ghosh, Sanjay Kumar and Toshniwal, Durga and Sekimoto, Yoshihide}, journal={Geoscience Data Journal}, volume={11}, number={4}, pages={846--862}, year={2024}, publisher={Wiley Online Library} }@article{gao2025ecemyolo, title={Enhanced Road Infrastructure Defect Detection via Attention-Based Visual Inspection}, author={Ningbo Gao, Dupengfei Zhai, Guifang Shi, Guanya Hao, Yong Qi, Zi Yang}, journal={Journal of Real-Time Image Processing}, year={202X}, publisher={Springer}, note={Code: https://github.com/zdpf122/ECEM-YOLO-RoadDefect; Dataset: https://doi.org/10.5281/zenodo.17792244}}