BraDD-S1TS: Brazilian Amazon Deforestation Detection from Sentinel-1 Time Series (TACO)
- Provider
- Karaman et al. (University of Zurich); TACO repackaging by ISP, Universitat de Valencia
- License
- CC-BY-4.0 (source)
- Size
- 14 files
- Format
- PNG (in sampled files)
- Stored at
- Hugging Face (https://huggingface.co/datasets/isp-uv-es/bradd-s1ts-taco)
- Access
- Free and public. No account with the data provider is needed.
- DOI
- 10.5194/isprs-annals-X-1-W1-2023-835-2023
- Category
- Earth Observation & Geospatial
- Source
- https://huggingface.co/datasets/isp-uv-es/bradd-s1ts-taco
About this dataset
BraDD-S1TS (ISPRS Annals, 2023) is a change detection benchmark for deforestation in the Brazilian Amazon using radar only, because clearing happens in the wet season when optical sensors see cloud. It contains 25,988 Sentinel-1 GRD time series of 48x48 pixels with 10 to 65 acquisitions each, VV and VH sigma-nought backscatter in dB, with a binary deforestation mask for the observed window and a second mask of clearing that predates it. Samples follow the release's seven sampling strategies: one positive and six negatives (boundary, old deforestation, forest, herbaceous, agriculture, shrubs). The official closed-set split is 15,625 train, 5,112 validation and 5,251 test. No CRS or coordinates are published for the patches; only the Brazilian state is given. About 16.6 GB in TACO format, read with tacoreader. Original on Zenodo under CC BY 4.0; cite Karaman et al. 2023.
How to load this dataset in Python
from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="isp-uv-es/bradd-s1ts-taco", filename="assets/logos.png", repo_type="dataset")
print(path)Data structure
File types in sampled files: PNG
Files
Example file: assets/logos.png
Analyze this dataset without downloading it
On Scigantic, BraDD-S1TS: Brazilian Amazon Deforestation Detection from Sentinel-1 Time Series (TACO) opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.
This page is also available as Markdown and as JSON.
Related Earth Observation & Geospatial datasets
- NASA Prediction of Worldwide Energy Resources (POWER)
- AIRS/Aqua L1C Infrared (IR) resampled and corrected radiances V6.7 (AIRICRAD) at GES DISC
- SPHEREx Quick Release (QR): An All-Sky Spectral Survey
- Hansen Global Forest Change 2000-2025 (v1.13): 30 m Landsat Tree Cover, Forest Loss Year and Gain Tiles
- Land Surface Temperature
- A2D2: Audi Autonomous Driving Dataset
- Plateau Tokyo Buildings in FlatCityBuf
- TreeSatAI Time Series: Tree Species Labels with Aerial Imagery and Yearly Sentinel-1/2 Series (Lower Saxony)