metadata

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-nc-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

License

This dataset is released under Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0)

You are free to:

Share: Copy and redistribute the material in any medium or format
Adapt: Remix, transform, and build upon the material