README.md

---
license: cc-by-sa-4.0
size_categories:
- 1M<n<10M
task_categories:
- other
tags:
- earth-observation
- remote-sensing
- sentinel-1
- sar
- synthethic-aperture-radar
- satellite
- geospatial
dataset_info:
- config_name: default
  features:
  - name: product_id
    dtype: string
  - name: grid_cell
    dtype: string
  - name: product_datetime
    dtype: string
  - name: thumbnail
    dtype: image
  - name: vv
    dtype: binary
  - name: vh
    dtype: binary
configs:
- config_name: default
  data_files: images/*.parquet
- config_name: metadata
  data_files: metadata.parquet
---

# Core-S1RTC

Contains a global coverage of Sentinel-1 (RTC) patches, each of size 1,068 x 1,068 pixels. This dataset is part of the **Major TOM** ecosystem.

**Papers:**
- [Major TOM: Expandable Datasets for Earth Observation](https://huggingface.co/papers/2402.12095)
- [EarthEmbeddingExplorer: A Web Application for Cross-Modal Retrieval of Global Satellite Images](https://huggingface.co/papers/2603.29441)

**Links:**
- **Project Page:** [EarthEmbeddingExplorer](https://modelscope.ai/studios/Major-TOM/EarthEmbeddingExplorer)
- **GitHub Repository:** [ESA-PhiLab/Major-TOM](https://github.com/ESA-PhiLab/Major-TOM)

| Source | Sensing Type | Number of Patches | Patch Size | Total Pixels |
|--------|--------------|-------------------|------------|--------------|
|Sentinel-1 RTC | Synthetic Aperture Radar |1,469,955|1,068 x 1,068 (10 m) | > 1.676 Trillion |

## Content

| Column | Details | Resolution |
|--------|---------|------------|
| VV | Received Linear Power in the VV Polarization | 10m | 
| VH | Received Linear Power in the VV Polarization | 10m | 
| thumbnail | Rescaled false colour<sup>1</sup> saved as png | 10m | 


<sup>1</sup> False colour composites are made with decibel-scale values with red green and blue defined as ```R:VV  G:VV+VH  B:VH```. For each channel, a contrast-stretch is applied, transforming minimum-maximum to 0-255.  This means bluer areas have relatively higher VH values, whilst brightness is a function of overall intensity. This is relative within each thumbnail because of the normalisation, and so cannot be compared across different samples.

## Spatial Coverage
This is a global monotemporal dataset. Nearly every piece of Earth captured by Sentinel-1 is contained at least once in this dataset (and only once, excluding some marginal overlaps). The coverage is about 35% lower than for Core Sentinel-2 dataset due to the sensor coverage limitations.

The following figure demonstrates the spatial coverage (only black pixels are absent):
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6304c06eeb6d777a838eab63/hFf98OyO25mL9Q962d5NL.png)

## Example Use

Interface scripts are available at https://github.com/ESA-PhiLab/Major-TOM

Here's a sneak peek with a thumbnail image:
```python
from fsspec.parquet import open_parquet_file
import pyarrow.parquet as pq
from io import BytesIO
from PIL import Image

PARQUET_FILE = 'part_03900' # parquet number
ROW_INDEX = 42 # row number (about 500 per parquet)

url = "https://huggingface.co/datasets/Major-TOM/Core-S1RTC/resolve/main/images/{}.parquet".format(PARQUET_FILE)
with open_parquet_file(url,columns = ["thumbnail"]) as f:
    with pq.ParquetFile(f) as pf:
        first_row_group = pf.read_row_group(ROW_INDEX, columns=['thumbnail'])

stream = BytesIO(first_row_group['thumbnail'][0].as_py())
image = Image.open(stream)
```

## Cite
[![arxiv](https://img.shields.io/badge/Open_Access-arxiv:2402.12095-b31b1b)](https://arxiv.org/abs/2402.12095/)
```latex
@inproceedings{Major_TOM,
  title={Major TOM: Expandable Datasets for Earth Observation}, 
  author={Alistair Francis and Mikolaj Czerkawski},
  year={2024},
  booktitle={IGARSS 2024 - 2024 IEEE International Geoscience and Remote Sensing Symposium}, 
  eprint={2402.12095},
  archivePrefix={arXiv},
  primaryClass={cs.CV}
}
```

Powered by [Φ-lab, European Space Agency (ESA) 🛰️](https://huggingface.co/ESA-philab)