Downloads · 30 days
0
rosielickorish/ibm-geospatial-saltmarsh-uk-rgbdem-extent-2m
ibm-geospatial-saltmarsh-uk-rgbdem-extent-2m is a machine learning model from rosielickorish. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as openmdw-1.0.
This is a fine-tuned geospatial foundation model for detecting saltmarsh habitat in England using multi-spectral satellite imagery.
Downloads · 30 days
0
Access
Public
Updated Jun 22, 2026
Repo size
362 MB
Likes
0
Public
Click a slice to open those files.
.ckpt362 MB · 100%
From the Hugging Face model README
This is a fine-tuned geospatial foundation model for detecting saltmarsh habitat in England using multi-spectral satellite imagery.
The model identifies the extent of saltmarsh habitats from aerial imagery and was trained on labelled English saltmarsh extents from the Environment Agency. The model was fine-tuned from TerraMind, a general-purpose geospatial foundation model developed by ESA and IBM.
This model was trained using:
The extent and zonation of saltmarsh in England: 2016–2019
Environment Agency (2022)
https://www.gov.uk/government/publications/the-extent-and-zonation-of-saltmarsh-in-england-2016-2019
The model was trained on 256x256 labelled tiles, extracted at 10m resolution, where all saltmarsh zones (Pioneer, Spartina, Mid-Low, Upper, Reedbeds and Unclassified marsh) were merged to denote marsh extent.
The final training dataset size was train: 3539, val: 1156, test: 461 (total: 5156).
Saltmarshes in North West England were held out as the test set.
Vertical Aerial Photography
Environment Agency (2025)
https://www.data.gov.uk/dataset/4921f8a1-d47e-458b-873b-2a489b1c8165/vertical-aerial-photography
The model was trained using 3 bands: RGB
LIDAR Composite Digital Terrain Model
Environment Agency (2025)
https://www.data.gov.uk/dataset/e529ca2f-b4ce-403e-8cef-ab821061c4f3/lidar-composite-digital-terrain-model-dtm-2m
Elevation was provided at 2m resolution as an additional modality, as supported by Terramind.
Example workflows and tutorials are available via the ML4EO workshop repository:
👉 https://github.com/IBM/ML4EO-workshop-2026
This work was supported by the Hartree National Centre for Digital Innovation, a collaboration between STFC and IBM.
Anne Jones, Helen Tamura-Wicks, Rosie Lickorish & Geoffrey Dawson.
All content in this repository including code has been provided by IBM under the associated open source software license and IBM is under no obligation to provide enhancements, updates, or support. IBM developers produced this code as an open source project (not as an IBM product), and IBM makes no assertions as to the level of quality nor security, and will not be maintaining this code going forward.
© Environment Agency copyright and/or database right 2023. Licensed under the Open Government Licence v3.0.