Programmer-RD-AI/road-issues-detection-dataset

Road Issues Detection Dataset

Dataset Summary

This comprehensive dataset contains 9,660 high-resolution RGB images categorized for road infrastructure issues detection. The dataset focuses on identifying critical urban infrastructure problems including potholes, damaged roads, broken road signs, illegal parking violations, and environmental cleanliness issues. It has been specifically organized and curated for computer vision and machine learning applications in smart city infrastructure monitoring.

Supported Tasks and Leaderboards

• Image Classification: Classify different types of road infrastructure issues
• Object Detection: Detect and locate specific road problems within images
• Multi-label Classification: Images may contain multiple types of issues
• Urban Infrastructure Monitoring: Real-world application in smart city systems

Languages

The dataset primarily contains images with English text on road signs and infrastructure, though the visual nature makes it applicable across different linguistic contexts.

Dataset Structure

data/
├── Road Issues/
│   ├── Broken Road Sign Issues/     # Images of damaged or vandalized road signs
│   ├── Damaged Road issues/         # Images of broken roads and speed breakers  
│   ├── Pothole Issues/             # Images of potholes and road surface damage
│   ├── Illegal Parking Issues/     # Images of illegal parking violations
│   └── Mixed Issues/               # Images containing multiple types of issues
└── Public Cleanliness + Environmental Issues/
    ├── Littering Garbage on Public Places Issues/  # Environmental and cleanliness issues
    └── Vandalism Issues/                           # Images of graffiti and vandalism

Dataset Creation

Source Data

This dataset was compiled from multiple high-quality sources specializing in urban infrastructure monitoring:

1. Litter Street Images: Roboflow Universe - Kabml Images
2. Urban Trash Dataset: Kaggle - Domestic Trash Garbage Dataset
3. Road Issues Plot Holes Pavement: Roboflow Universe - Navrachana University
4. Graffiti and Vandalism Dataset: Roboflow Universe - Hruts Workspace
5. Damaged Road Signs Dataset: Roboflow Universe - Jayke Boghean
6. Damaged Construction Dataset: Mendeley Data
7. Illegal Parking Dataset: Roboflow Universe - Parking AMU50
8. Additional Damaged Road Signs: Roboflow Universe - Road Inspection

Data Collection and Processing

The images were carefully curated, organized, and categorized to ensure high quality and relevance for infrastructure monitoring applications. All images are in JPEG format with consistent quality standards suitable for machine learning model training and evaluation.

Categories and Use Cases

Road Issues

• Broken Road Sign Issues: Contains images of graffitied, damaged, or vandalized road signs for automated sign maintenance systems
• Damaged Road Issues: Images showing road surface damage, speed breakers, and general road infrastructure problems for municipal planning
• Pothole Issues: Specifically focused on pothole detection and road surface deterioration for automated road quality assessment
• Illegal Parking Issues: Documentation of parking violations and improper vehicle placement for traffic management systems
• Mixed Issues: Images that contain multiple types of road-related problems for comprehensive monitoring

Public Cleanliness + Environmental Issues

• Littering Garbage on Public Places Issues: Environmental cleanliness monitoring for smart city applications
• Vandalism Issues: Images of graffiti and vandalism for public space maintenance

Dataset Statistics

• Total Images: 9,660 JPG images
• Format: RGB images in JPEG format (.jpg)
• Categories: 7 distinct issue types across infrastructure and environmental problems
• Organization: Hierarchical structure with clear category separation
• Quality: High-resolution images suitable for machine learning applications
• File Size: Varies by image, optimized for computer vision applications

Applications

This dataset is particularly suitable for:

• Road Infrastructure Monitoring Systems: Automated detection and classification of road problems
• Municipal Maintenance Planning: Data-driven infrastructure maintenance scheduling
• Smart City Infrastructure Management: Integration with IoT and smart city platforms
• Computer Vision Research: Academic and commercial research in urban infrastructure
• Machine Learning Model Training: Training robust models for real-world deployment
• Automated Inspection Systems: Integration with drone or vehicle-mounted camera systems

Licensing Information

This dataset is released under the CC0-1.0 (Creative Commons Zero) license, making it freely available for any use, including commercial applications, without any restrictions.

Citation Information

If you use this dataset in your research or applications, please cite:

@misc{ranuga_disansa_2025,
    author       = { Ranuga Disansa and Rivindu Ashinsa and Sanila Wijesekara and Thuan Naheem },
    title        = { road-issues-detection-dataset (Revision c0c9c37) },
    year         = 2025,
    url          = { https://huggingface.co/datasets/Programmer-RD-AI/road-issues-detection-dataset },
    doi          = { 10.57967/hf/6188 },
    publisher    = { Hugging Face }
}

Contributions

This dataset was created and curated by the OutlierRejects team:

• Ranuga Disansa
• Rivindu Ashinsa
• Sanila Wijesekara
• Thuan Naheem

We acknowledge and thank all the original data sources and their contributors for making their datasets available to the community.

Additional Information

Dataset Curators

OutlierRejects Team

Dataset Version

Current version: c0c9c37 (2025)

Contact Information

For questions, issues, or contributions, please refer to the dataset repository or contact the curators through the Hugging Face dataset page.