OPR-Project/ITLP-Campus-Indoor

🏢 ITLP Campus Indoor is a multimodal dataset focused on indoor Place Recognition across five floors of a university building. It features synchronized RGB images from front and back cameras, LiDAR point clouds, manually annotated scene text, and strategically placed ArUco markers for accurate localization. Semantic segmentation masks were automatically generated using the OneFormer model. Captured during night and twilight conditions, the dataset reflects real-world challenges in indoor navigation, including visually similar corridors and repetitive layouts. It’s ideal for research on text-aware, vision-based, and multimodal localization in complex indoor environments.

This dataset can be used for:

• 🧪 Developing and testing localization algorithms using real-world data collected across various seasons, times of day, and weather conditions.
• 📚 Educational and research projects on multimodal localization, machine learning, and computer vision.
• 📈 Benchmarking and comparative analysis of global localization algorithms.
• 🎯 Creating machine learning models robust to environmental changes and external conditions.

---

❗ Please note

This dataset is currently not compatible with the datasets library. The recommended way to download the data is by using the huggingface_hub library. Example code snippet:

from pathlib import Path
from huggingface_hub import snapshot_download

out_dir = Path("/dir/to/save/data")
out_dir.mkdir(parents=True, exist_ok=True)

snapshot_download(repo_id="OPR-Project/ITLP-Campus-Indoor", repo_type="dataset", local_dir=out_dir)

For reading and working with the data, we recommend using the OpenPlaceRecognition library: https://github.com/OPR-Project/OpenPlaceRecognition