---
license: mit
---

# Dataset: PhilEO Downstream Tasks

A novel 400GB Sentinel-2 dataset of the PhilEO Bench containing labels for the three downstream tasks of building density estimation, road segmentation, and land cover classification.

## Dataset Details

### Dataset Description

The PhilEO dataset is a 400GB global dataset of Sentinel-2 images and has labels for roads, buildings, and land cover, where these are the three downstream tasks. The data is sampled from geographically diverse regions around the globe including: Denmark, East Africa, Egypt, Guinea, Europe, Ghana, Israel, Japan, Nigeria, North America, Senegal, South America, Tanzania, and Uganda. Each region has up to 200 tiles of varying sizes. Some locations have been revisited up to 3 times.

The data contain 11 bands at 10m resolution in the following order: 0-SCL, 1-B02, 2-B03, 3-B04, 4-B08, 5-B05, 6-B06, 7-B07, 8-B8A, 9-B11, and 10-B12 where SCL is the Scene Classification Layer.

- **Curated by:** ESA Phi-lab
- **License:** MIT

## Uses

The dataset can be used to evaluate any EO Foundation Model.

### Dataset Sources

The basic links for the dataset:

- **Repository:** http://huggingface.co/datasets/ESA-philab/PhilEO-downstream 
- **Paper:** http://arxiv.org/pdf/2401.04464.pdf 
- **Project Website:** http://phileo-bench.github.io 
- **Code GitHub:** http://github.com/ESA-PhiLab/PhilEO-Bench 
- **Dataset also in:** http://www.eotdl.com/datasets/PhilEO-downstream 
- **arXiv:** http://arxiv.org/abs/2401.04464 

## Citation

Casper Fibaek, Luke Camilleri, Andreas Luyts, Nikolaos Dionelis, and Bertrand Le Saux, “PhilEO Bench: Evaluating Geo-Spatial Foundation Models,” arXiv:2401.04464, 2024.