Standards

STAC Updates, February 2022

In the last update post, I shared that we had a great set of sponsors come in to further the greater STAC ecosystem. I’m pleased to share that we had a first successful ‘funders call’ to determine the priorities for the money we raised. This probably slowed the process a bit, but it was super valuable for the STAC Project Steering Committee to hear from key sponsors what they see as most needed in the STAC ecosystem. It was a great conversation, and I look forward to future calls.

Machine Learning

Detecting Agricultural Croplands from Sentinel-2 Satellite Imagery

A guide to identifying croplands with reasonable accuracy using a semantic segmentation model. We developed UNet-Agri, a benchmark machine learning model that classifies croplands using open-access Sentinel-2 imagery at 10m spatial resolution with ground reference data provided by the Western Cape Department of Agriculture in South Africa. This post is a step-by-step walkthrough of how we developed the model and evaluated its performance. Understanding what UNet-Agri does will help you build your model and deploy it for a similar application.

News

Winning Solutions to Prediction Crop Types in Western Cape, South Africa from Data Challenge

In this webinar, the first-place winners of the Radiant Earth Spot the Crop data challenge on Zindi present their winning solutions to prediction crop types in Western Cape, South Africa. Raphael Kiminya from Kenya won the track that used satellite image time-series of Sentinel-2 multi-spectral data as input to his model. MG Ferreira from South Africa and Tien-Dung LE from Belgium teamed up to win the track that used Sentinel 2 and Sentinel-1 (radar) satellite data as input to their model. Mike Wallace from the Western Cape Department of Agriculture and Phillip Olbrich from GIZ will provide brief remarks, followed by a Q&A session. Watch the recording.

News

Nominations Open for the 2022 Radiant MLHub Impact Award

The call for nominations is open to any individual or team building agricultural-related applications for Africa. The Radiant MLHub Impact Award is a new annual prize organized by Radiant Earth Foundation. The award will recognize an individual or team contributing to real-world applications that address agriculture issues with geospatial training data found on Radiant MLHub. Launched in 2019, Radiant MLHub is an open library dedicated to Earth observation training data and machine learning models.

The Radiant MLHub Impact Award aims to raise awareness about geospatial data to tackle the urgent challenges of climate change, promote local solutions to local problems, and facilitate greater collaboration across organizations and individual practitioners.

Community Voices

Radiant MLHub Spotlight Q&A: Renate Thiede

Combining mathematical statistics, geospatial data, and artificial intelligence in support of global development.

Renate is an alumnus of Radiant Earth’s first virtual ML4EO training of trainers bootcamp that focused on using machine learning with satellite data. The bootcamp ran from May 3–14, 2021, thanks to a grant from the GIZ FAIR Forward- Artificial Intelligence for all program, which the Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) implements on behalf of the German Federal Ministry for Economic Cooperation and Development (BMZ).

In this Q&A, we sat down with Renate to discuss her journey combining statistics, geospatial data, and AI. She also shares her experience participating in the ML4EO Bootcamp.

Community Voices

Data Challenge Winners: Q&A with MG Ferreira and Tien-Dung LE

A conversation with the First Place winners of the Radiant Earth Spot the Crop XL Data Challenge. We recently announced the Radiant Earth Spot the Crop Data Challenge winners to predict crop types in Western Cape, South Africa using satellite image time-series. The competition was organized in two parallel tracks: In track 1, participants used time-series of Sentinel-2 multispectral data as input to their model; In track 2, both Sentinel-2 and Sentinel-1 (radar) data were required as input.

In this Q&A, we sat down with MG Ferreira from South Africa and Tien-Dung LE from Belgium to talk about their journey to becoming data scientists and their approach to tackling the problem. MG and Tien-Dung won the Spot the Crop XL Data Challenge winners that used Sentinel-1 radar and Sentinel-2 multispectral data as input to the model.

Community Voices

Data Challenge Winner: Q&A with Raphael Kiminya

A conversation with the First Place winners of the Radiant Earth Spot the Crop Data Challenge.
We recently announced the Radiant Earth Spot the Crop Data Challenge winners to predict crop types in Western Cape, South Africa using satellite image time-series. The competition was organized in two parallel tracks: In track 1, participants used time-series of Sentinel-2 multispectral data as input to their model; In track 2, both Sentinel-2 and Sentinel-1 (radar) data were required as input.

In this Q&A, we sat down with Raphael Kiminya from Kenya to talk about his journey to becoming a data scientist and his approach to tackling the problem. He won the Spot the Crop Data Challenge that used Sentinel-2 multispectral data as input to the model.

Standards

The Exciting Future of the STAC Browser

More and more organizations benefit from the STAC specification and publish their geospatial data with STAC metadata. Having the data published in a standardized way in JSON is paving the way to wider adoption of geospatial data so that humans can find and make good use of the data. There is no doubt having JSON files is nice for developers, but it is equally important to spread the data to professionals in other sectors. Those users often need an easily accessible and searchable graphical user interface for the data. That is where STAC Browser comes into play.

Machine Learning, News

Geospatial Models Now Available in Radiant MLHub

The models include metadata based on the STAC ML Model Extension to enable easy sharing and retrieval. Radiant MLHub has been the source for high-quality open geospatial training data for use with machine learning (ML) algorithms since 2019. Today, we’re excited to announce the addition of a model repository allowing Radiant MLHub users access to both geospatial training data and ML models. The geospatial models catalog includes metadata that describes training data associated with a model and its architecture for training a model to generate predictions.

Machine Learning, News

Available Now: Machine Learning for Earth Observation Online Course

We have the pleasure of introducing Radiant Earth Foundation’s first online course, Machine Learning for Earth Observations (ML4EO) Bootcamp. Available on Atingi, an open digital learning platform designed to improve training and employment opportunities, this self-paced course contains a mixture of lectures and hands-on exercises for novice data science or remote sensing practitioners. Atingi is implemented by the Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) on behalf of the German Federal Ministry for Economic Cooperation and Development (BMZ).