Taylor Geospatial Engine

Want to make waves in the geospatial world & learn from industry leaders? ?? https://go.aws/4eWQvM0 Join #AWS & Taylor Geospatial Institute in St. Louis for the kickoff of an exclusive geospatial data & #generativeAI challenge! ???? Innovators in North America will be invited to submit proposals showcasing scalable applications. ?? Oct 29th from 9 AM - 5 PM ?? Learn from industry experts ?? Compete for $1M in AWS credits Don't miss out!

Thanks to Alex Mu?oz Wilson for capturing this candid shot (which I asked for ??) of me talking about **Fields of the World** in today's lecture in my Planetary-Scale Data Institutions course at Yale Jackson School of Global Affairs and Yale School of the Environment. Fields of the World is a shining example of the kind of ambitious and transformational planetary-scale machine learning data products that are possible today. I'm very proud that Radiant Earth has been able to collaborate with Taylor Geospatial Engine to support it. ? Hannah Kerner, Nathan Jacobs, Jennifer Marcus, Chris Holmes, Matthias Mohr ?

Taking a moment to celebrate accomplishments is crucial -- and there were 2 big ones for us this week! (More info here: https://lnkd.in/gtqxKDph) The team of collaborators in our Innovation Bridge initiative released the largest training dataset for AI models to extract field boundaries from satellite imagery: Fields of the World (https://fieldsofthe.world)! And the team brought together by CSIS released the 2024 Commercial Remote Sensing Global Rankings (https://lnkd.in/gTUdw8g5). We appreciate all of the good work being done in these communities and are simply thrilled to contribute. Great thanks to: Hannah Kerner, Caleb Robinson, Nathan Jacobs, Center for Strategic and International Studies (CSIS), United States Geospatial Intelligence Foundation (USGIF), Taylor Geospatial Institute, Radiant Earth

?? Fields of The World (FTW) is live! ?? FTW is the most diverse and comprehensive machine learning dataset for field instance segmentation in satellite images. With 70,000+ samples from 24 countries across 4 continents, FTW aims to accelerate advancements in AI for agricultural and environmental monitoring. ???? We invite the community to contribute and help us expand the dataset further. This open-source project combines semantic and instance labels with Sentinel-2 RGB-NIR data across multiple dates for even more robust analyses. Interested in joining or exploring? Find all the resources below! ?? ?? Paper: https://lnkd.in/ebvJuYKv ?? Website: https://fieldsofthe.world/ ?? Data: https://lnkd.in/exQZqd9z ?? Code: https://lnkd.in/erJEjknP #AI #MachineLearning #GeospatialData #Agriculture #OpenSource #RemoteSensing #DataScience

?? Thrilled to Announce: Fields of The World (FTW)! ?? I’m excited to share FTW, a groundbreaking benchmark dataset for agricultural field boundaries??????. As the largest and most diverse dataset for field instance segmentation in satellite imagery, FTW offers:?? ? Over 70,000 samples from 24 countries across four continents. ? Comprehensive semantic and instance labels, along with Sentinel-2 RGB-NIR data from two different dates. ? Deep learning models pre-trained on FTW achieve better transfer learning results than those trained on more geographically limited datasets (e.g., Europe only). ? Remarkably, FTW models showcase strong zero-shot capabilities, enabling segmentation in areas without prior training data. Discover more about FTW: ?? FTW Website: https://fieldsofthe.world/ ?? Research Paper: https://lnkd.in/ebvJuYKv ?? Data Access: https://lnkd.in/exQZqd9z ?? GitHub Repository: https://lnkd.in/erJEjknP It was an awesome experience working with Hannah Kerner, Snehal Chaudhari, Aninda Ghosh, Caleb Robinson, Eddie Choi, Nathan Jacobs, Chris Holmes, Matthias Mohr, Rahul Dodhia, Juan M. Lavista Ferres, and Jennifer Marcus #AI #MachineLearning #GeospatialData #Agriculture #OpenSource #RemoteSensing #DataScience #FieldBoudaries

Today, a team of four institutions, including the Taylor Geospatial Institute, Taylor Geospatial Engine, the United States Geospatial Intelligence Foundation (USGIF), and the Center for Strategic and International Studies (CSIS) unveiled the 2024 Commercial Remote Sensing Global Rankings, a report that identifies a “Top 3” of the world’s best commercial space-based remote sensing systems in various categories. Download the full report and hear an audio summary here - https://lnkd.in/eWq2yq-D

I encourage you to attend one of our Info Sessions if you are considering a PhD program in Computer Science, Imaging Science, Data Science, Electrical & Systems Engineering, or other areas in Engineering. You'll learn a lot about our programs and be eligible for an application fee waiver. See https://lnkd.in/g52CZy4i to register and for additional details.

Little update on Source Cooperative: - Now hosting over 400TB of data published by 64 organizations or individuals - Served up 6.9TB of data last week, across 6.4 million GET requests (this is a remarkably big week! Our 52-week rolling average is .75TB and 3.8 million GETs) - data.source.coop URLs show up in 152 results on GitHub search (link in comments) And we're still in beta! About to step on the gas. ??

We're now the Cloud-Native Geospatial Forum (CNG)! Learn more about our plans to connect and expand the geospatial data user community: https://lnkd.in/eYsquuuA

We are excited to share another technical update from?Matthias Mohr?and the?Cloud-Native Geospatial Foundation?on the development and use of open source tools to accelerate innovation in AI and computer vision models to extract field boundaries from earth observation imagery. Read more here - https://lnkd.in/ge_XfDvm