Robin Cole

Excited to share that the first paper of my PhD has been accepted for publication at the ISPRS Geospatial Week 2025! This dataset paper introduces a globally representative, high-resolution (10m) benchmark dataset for Above Ground Biomass estimation.

Deepness QGIS plugin With Marek Kraft

This is great stuff from Robin Cole and Marek Kraft

www.satellite-image-deep-learning.com/p/deepness-q...

Thanks for sharing 🙏

Prithvi-EO-2.0: A Versatile Multi-Temporal Foundation Model for Earth Observation Applications

github.com/NASA-IMPACT/...

🌍 How can AI transform hurricane forecasting?

NASA’s Prithvi AI model, built on decades of open data, is helping scientists uncover new insights about hurricanes while making forecasting more accessible to all.

🔗 Listen: www.nasa.gov/podcasts/cur...

#OpenScience #ClimateTech #DisasterResponse

Can you DM me your email?

The bad dads are the ones that don’t bother..!

New AI tool generates realistic satellite images of future flooding With help from AI, MIT scientists developed a method that generates satellite imagery from the future to depict how a region would look after a potential flooding event.

news.mit.edu/2024/new-ai-... #space #satellite #algorithm #disaster #hydrology #GAN

The EarthDaily platform will be public access soon

People are using Google study software to make AI podcasts—and they’re weird and amazing NotebookLM is a surprise hit. Here are some of the ways people are using it.

AI may sound more “natural” than how we, many of us, scientists speak.. am I too cynical to think that we could actually learn from this www.technologyreview.com/2024/10/03/1...

Here is an example of semantic segmentation of man-made structures (suburban housing in this case) with a FCN using UNET architecture. The houses missed were built after the GEE image was archived. See my shared COLAB notebook at

colab.research.google.com/drive/136YAV...

for the details.

TorchSpatial offers a comprehensive framework and benchmark suite designed to advance Spatial Representation Learning (SRL).

github.com/seai-lab/Tor...

Are geospatial foundation models actually useful? 🤔
Rolf et al. find a small MLP trained on curated local lidar data to outperform published tree canopy height maps & fine-tuned foundation models in a case study in south-eastern Africa. @estherrolf.bsky.social
arxiv.org/html/2411.14...

Proper release Friday when we merge docs, but if you can't wait:
1. Clay v1.5 Largest #AIforEarth model, works with any satellite. huggingface.co/made-with-cl...
2. Also embedded the ENTIRE naip archive. 10b embeddings
source.coop/repositories...
All open source, open data, open license. #satsummit

I've initiated a starter pack with researchers working on Deep Learning (SSL, generative models, foundation models, deep learning at scale...) for Earth observation/remote sensing that I know/that I've found on this network!

-> Don't hesitate to ping me to be added/removed! 🌏

go.bsky.app/EKVwvde

Hello world 👋