PCFootprint: A Large-scale Benchmark for Vectorized Building Footprint Extraction from ALS Point Clouds

PCFootprint introduces the first extensive, open-source benchmark specifically engineered for the vectorized extraction of building footprints from ALS point clouds. Sourced from the Estonian Land and Spatial Development Board, the dataset spans a vast geographic extent of 540.67 km², comprising 33,000 standardized tiles and 227,264 meticulously annotated building instances across diverse urban and rural landscapes.

📊 Dataset Specifications

The detailed specifications of the PCFootprint dataset are summarized below:

Property	Details
Subsets	Mainland, Islands
Data Type	ALS Point Cloud
Annotation Format	Vectorized Polygon
Accessibility	Open-source
Point Spacing [m]	0.22 – 1.09
Tile Size [m]	128 × 128
Total Tiles	33,000
Building Instances	227,264
Coverage [km²]	540.67

---

📂 Data Structure

The dataset is organized into two primary subsets representing different geographic characteristics of Estonia:

PCFootprint/
├── Mainland/
│   ├── pointcloud/    # Point cloud data (e.g., .laz)
│   └── footprint/     # Vectorized footprint annotations (e.g., .json)
└── Islands/
    ├── pointcloud/
    └── footprint/

---

⚖️ License & Acknowledgement

Dataset License

The PCFootprint dataset is licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0).

Data Origin & Source Attribution

The development of this dataset was made possible through the open data provided by the Republic of Estonia Land and Spatial Development Board. We gratefully acknowledge the following data sources used in the creation of PCFootprint:

• Elevation / Raw LiDAR Data: (2021-2024)
• Administrative Units / County Data: (2025)
• Topographic Database (ETAK) / Buildings: (2021-2024)

All original data are used in accordance with the Estonian Land Board Open Data Licence (01.01.2025).