OpenConstruction: A Systematic Synthesis of Open Visual Datasets for Data-Centric Artificial Intelligence in Construction Monitoring

🌐 Quick Links

• 💥 New Website: https://www.openconstruction.org/
• Suggest a New Dataset: Submit here
• Share an Idea or Feedback: Open suggestion
• Request Dataset Removal: Contact us at support@openconstruction.org from a verifiable email

📋 Overview

OpenConstruction dataset catalog is a curated collection of open-access datasets for construction monitoring and analysis. It provides a centralized resource for researchers, practitioners, and developers working on AI applications in the construction industry.

🎯 Motivation

AI-driven computer vision is transforming construction monitoring, safety management, and productivity. Yet, robust AI development requires access to diverse and well-documented datasets.

This project addresses that need by:

• Compiling a comprehensive catalog of construction datasets
• Standardizing metadata for easier comparison and selection
• Supporting benchmarking and reproducible research
• Facilitating AI model development across tasks and modalities

📊 Dataset Categories

Our catalog spans a variety of construction-relevant tasks:

• Object Detection
• Semantic Segmentation
• Action Recognition
• Pose Estimation
• Image Captioning
• SLAM / Visual Localization

📈 Dataset Statistics

• Modalities: RGB, Thermal, Depth, LiDAR/Point Cloud, Synthetic
• Annotations: Bounding boxes, segmentation masks, keypoints, captions
• Geographic Coverage: Multi-country, diverse contexts

🤝 Contributing

We welcome community contributions to grow and improve this catalog. Use the quick templates:

• Suggest a New Dataset
• Share Feedback

👥 Community Contributors

This project is enriched by the support and contributions from the community. Thank you for helping grow OpenConstruction!

See full contributor list here: OpenConstruction Community

📧 Contact

For questions, suggestions, or collaboration:

• Open an issue
• Email: support@openconstruction.org

🙏 Acknowledgments

We thank all dataset creators and contributors for making their resources publicly available to advance AI in construction.

Reference

@misc{xiong2025openconstructionsystematicsynthesisopen,
      title={OpenConstruction: A Systematic Synthesis of Open Visual Datasets for Data-Centric Artificial Intelligence in Construction Monitoring}, 
      author={Ruoxin Xiong and Yanyu Wang and Jiannan Cai and Kaijian Liu and Yuansheng Zhu and Pingbo Tang and Nora El-Gohary},
      year={2025},
      eprint={2508.11482},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2508.11482}, 
}