MMDC Europe (full dataset)
Authors/Creators
Description
MMDC-EU is a multi-year, multimodal SITS dataset spanning Europe. This dataset is composed of the following data: the Sentinel 2 L2A product (processed with Sen2cor), the Sentinel 1 ascending and descending orbits, ECMWF AGERA5 weather variables, and a digital elevation model (DEM). Although all modalities are not initially available at a 10m resolution, all products are up-sampled on the Sentinel-2 10m spatial grid. Multi-year multimodal SITS were acquired from January 2017 to December 2020.
| Product name | Bands | Average acquisitions/year |
| S2 L2A | B2-B8, B8A, B11,B12,SCL, CLM, dataMask, sunAzimuthAngles,sunZenithAngles | 89 |
| S1-GRD-\gamma 0 terrain ASCENDING | VV, VH | 90 |
| S1-GRD-\gamma 0 terrain DESCENDING | VV, VH | 107 |
| ERA 5 | Dewpoint-temperature, Precipitation-flux, solar-radiation-flux, temperature-max, temperature-mean, temperature-min, vapour-pressure, wind-speed | 365 |
| DEM | GLO-30 | - |
Specifically, the pre-training dataset gathers data from eighteen S2 tiles. This pre-training dataset is divided into training and validation sets. To build the training dataset, ten smaller non-overlapping ROIs of size (512,512) are randomly selected from each of the twelve training tile. The disjoint validation dataset is composed of the remaining six S2 tiles, from which thirty ROIs of size (128,128) are randomly drawn. The code utilized to create the multimodal dataset is provided, allowing for potential future expansion : https://src.koda.cnrs.fr/iris.dumeur/openeo_datasets
The training data are available:
- 30TXT: 10.5281/zenodo.12750451
- 31TDL: 10.5281/zenodo.12755396
- 31TEN: 10.5281/zenodo.12773938
- 32TPT: 10.5281/zenodo.12795519
- 32TPQ: 10.5281/zenodo.1278467
- 32UMB: 10.5281/zenodo.13313224
- 33TVM: 10.5281/zenodo.13384485
- 32UQD: 10.5281/zenodo.1332334
- 33UXR: 10.5281/zenodo.13482459
- 33TYJ: 10.5281/zenodo.13385645
- 34UEC: 10.5281/zenodo.13628736
- 34TFR: 10.5281/zenodo.13618543
The validation data are available at:
- 30TYR: 10.5281/zenodo.13629634
- 31TEK: 10.5281/zenodo.13629668
- 32TNR: 10.5281/zenodo.13643024
- 32UPC: 10.5281/zenodo.13644104
- 33TXK: 10.5281/zenodo.13644596
- 34UDB: 10.5281/zenodo.13645759
The geojson describing the ROI extent on each S2 tiles are also provided in this repository.
Files
MMDCEU_description.md
Files
(104.4 kB)
| Name | Size | Download all |
|---|---|---|
|
md5:0fe053d3f3c095890049663720ba9181
|
2.0 kB | Preview Download |
|
md5:276092a3f17e2503ff6c3835e3a70497
|
3.5 kB | Download |
|
md5:0afd4a57673dfc68952ed6d35c4cece0
|
3.4 kB | Download |
|
md5:6226a995428a279ef2cd543c929c156e
|
3.4 kB | Download |
|
md5:629f31c586d5ee8fb8fc7bfe94abc5c5
|
3.5 kB | Download |
|
md5:6d8ff507899afc83abfa624061aacb13
|
3.5 kB | Download |
|
md5:6d917fc1b566e3cfdab242c026451cfc
|
3.4 kB | Download |
|
md5:ddec2ff4e4ee547bf1b21084ba0dd8ad
|
3.5 kB | Download |
|
md5:7487f5be8b91786ba3adca04689073c4
|
3.5 kB | Download |
|
md5:f23743529d7be76080dc98f7a01c7bc3
|
3.5 kB | Download |
|
md5:b91f09911cfab40703ff6da3b0405dad
|
3.5 kB | Download |
|
md5:0093fcd3f87e7c0cc2659af6060a591b
|
3.5 kB | Download |
|
md5:819823d4cac7cda4b51b7f9741be04bd
|
3.5 kB | Download |
|
md5:040b3be8c7a7cfd6f411b1ec0e12b895
|
10.1 kB | Download |
|
md5:1d38891eb7f51cfb85b9e8b251287c65
|
10.0 kB | Download |
|
md5:503aa7826955262a6d570c746c6fcea8
|
10.1 kB | Download |
|
md5:7f38ac74c971932961ab482d4beb9968
|
10.2 kB | Download |
|
md5:fa2bc5c0974472506a5b2f70dea29c74
|
10.2 kB | Download |
|
md5:fdff3363e25844dae95088915dfaa356
|
10.2 kB | Download |
Additional details
Related works
- Describes
- Preprint: arXiv:2407.08448 (arXiv)
- Is required by
- Software: https://src.koda.cnrs.fr/iris.dumeur/alise (URL)
- Is source of
- Software: https://src.koda.cnrs.fr/iris.dumeur/openeo_datasets (Other)
Funding
- Agence Nationale de la Recherche
- DeepChange - Deep generative models for detecting land cover changes from satellite image times series ANR-20-CE23-0003
- European Commission
- EVOLAND - EVOLUTION OF THE COPERNICUS LAND SERVICE PORTFOLIO INTEGRATING NOVEL EO DATA AND LATEST MACHINE LEARNING ALGORITHMS TO CONTINUOUSLY MONITOR THE STATUS, DYNAMICS AND BIOMASS OF THE LAND SURFACE 101082130