cindycui/GlobalGeoTree

---
language:
- en
pretty_name: GlobalGeoTree
size_categories:
- n>1M
tags:
- geospatial
- image
- webdataset
- biodiversity
- remote-sensing
- sentinel-2
- tree-species
- zero-shot-classification
- few-shot-classification
license: cc-by-4.0
---

# GlobalGeoTree Dataset

GlobalGeoTree is a comprehensive global dataset for tree species classification, comprising 6.3 million geolocated tree occurrences spanning 275 families, 2,734 genera, and 21,001 species across hierarchical taxonomic levels. Each sample is paired with Sentinel-2 image time series and 27 auxiliary environmental variables.

## Dataset Structure

This repository contains three main components:

### 1. GlobalGeoTree-6M
- Training dataset with around 6M samples
- Each sample includes:
- Sentinel-2 time series (12 monthly composites)
- 27 auxiliary environmental variables
- Hierarchical taxonomic labels (Family, Genus, Species)
- Format: WebDataset (.tar)

### 2. GlobalGeoTree-10kEval
- Evaluation dataset with carefully curated samples
- Three versions available:
- 90 species (30 each from Rare, Common, and Frequent categories)
- 300 species (100 each from Rare, Common, and Frequent categories)
- 900 species (300 each from Rare, Common, and Frequent categories)
- Format: WebDataset (.tar)

### 3. checkpoints
- Pre-trained GeoTreeCLIP model weights
- File: `GGT_6M.pth`
- Trained on the full GlobalGeoTree-6M dataset for 25 epochs

### 4. files
- Complete sample information file: `GlobalGeoTree.csv`
- Contains metadata for all samples including:
- Sample ID
- Taxonomic information (Family, Genus, Species)
- Geographic location (latitude, longitude)
- Source and year of observation
- Location description
- Format: CSV

## Related Repository

For detailed usage instructions, model implementation, and training scripts, please check our GitHub repository:
[GlobalGeoTree](https://github.com/MUYang99/GlobalGeoTree)

## License

This dataset is released under CC BY 4.0.

<!-- ## Citation

If you use this dataset in your research, please cite our paper:
```bibtex
@inproceedings{mu2025globalgeotree,
title={GlobalGeoTree: A Multi-Granular Vision-Language Dataset for Global Tree Species Classification},
author={Mu, Yang and Xiong, Zhitong and Wang, Yi and Shahzad, Muhammad and Essl, Franz and van Kleunen, Mark and Zhu, Xiao Xiang},
booktitle={Advances in Neural Information Processing Systems},
year={2025}
}
``` -->