In December 2019, we publicly launched Radiant MLHub, the first open-access cloud-based repository for geospatial training datasets. Since then, we have continuously published new datasets and expanded the ecosystem around Radiant MLHub.

The idea of Radiant MLHub was born in Spring 2018 after several discussions and feedback from members of the community and funders. We had started a new project to develop a global and geographically diverse land cover training dataset using human verification called LandCoverNet. Soon after the launch of LandCoverNet in 2018, we identified a gap in the ecosystem to facilitate publication and uptake of training datasets in our community. That gap in the data value chain led us to the design and implementation of Radiant MLHub.

Meet the 100+ organizations that focus on machine learning applications with satellite data

Building geospatial machine learning applications involve many dependable moving parts, from accessing Earth observation (EO) data, labeling imagery, and generating training data to creating and developing models and running analytics. A growing list of organizations from various sectors are providing solutions and services to advance these applications. Who can help you build machine learning applications, identify patterns from your data, or run your crowdsourcing campaign? What organizations are providing software or a platform that you can utilize to develop your machine learning model?

Jon Engelsman won the Best Quality Labeler award for our recent Data Labeling Contest. We asked him to detail his approach and workflow.

I recently attended the Cloud Native Geospatial Outreach Day, a virtual event designed to “introduce STAC, COG, and other emerging cloud-native geospatial formats and tools to new audiences.” As part of the outreach day, co-sponsors Planet, Microsoft, Azavea, and Radiant Earth teamed up to host a week-long data labeling contest. This friendly competition had contestants race to manually label the shapes of clouds across a large selection of satellite images from around the world. The contest’s ultimate goal was to generate a crowd-sourced collection of high-quality labeled images, data that can be used to train accurate cloud detection models.

Earlier this month, we organized a data labeling contest as part of the Cloud Native Geospatial Outreach Day sponsored by Planet, Microsoft, and Azavea. The contest was designed as a crowdsourcing campaign to encourage the global community to contribute to open-access training data catalogs. Participants were asked to identify cloudy pixels in Sentinel-2 scenes.

The labeling contest was conducted on GroundWork, Azavea’s annotation tool designed for geospatial data. We were amazed by the high participation worldwide and the community’s excitement to help develop a large-scale accurate cloud detection training dataset. In the end, 231 users around the world signed up and labeled 75,645 tasks, which equates to about 2 million km2 of classified Sentinel-2 imagery.

STAC specification is getting closer to the ver 1.0 milestone, and as such the first virtual Cloud Native Geospatial Sprint is being organized next week. An outreach day is planned on Sep 8th with a series of talks and tutorials for everyone. Read more about the sprint in this blog post by our Technology Fellow Chris Holmes. A new addition to the sprint is a data labeling contest!

If you have followed our blogs, we have written many times on the importance of open-access and high-quality labels on satellite imagery for building geospatial machine learning models. A scalable solution to generate labels for a large number of imagery is to run crowdsourcing campaigns and encourage the community to contribute to open-access training data catalogs.

Radiant Earth Foundation is pleased to announce receiving a 3-year cooperative agreement from NASA Advancing Collaborative Connections for Earth System Science (ACCESS) program to expand Radiant MLHub services, and support the growing demand for benchmarks in the geospatial machine learning community.

Radiant MLHub is the world’s first open-access cloud-based geospatial training data library designed to accelerate the adoption of ML to help solve global development challenges such as climate change and food insecurity.

LandCoverNet is an open-access global land cover classification training dataset with satellite image pixels labeled for seven land cover classes.

Radiant Earth Foundation is proud to announce the release of “LandCoverNet,” a human-labeled global land cover classification training dataset. Available for download on Radiant MLHub, the open geospatial library, LandCoverNet will enable accurate and regular land cover mapping allowing for timely insights into natural and anthropogenic impacts on the Earth. This release contains data across Africa, which accounts for ~1/5 of the global dataset.

Today, Radiant Earth Foundation announced the registration of a Spatio-Temporal Asset Catalog (STAC)-compliant version of SpaceNet’s high-quality geospatial labeled datasets for roads and buildings on Radiant MLHub. Radiant MLHub is the world’s first cloud-based open library dedicated to Earth observation training data for machine learning algorithms. The updated dataset catalog is also available through SpaceNet’s data registry.

Founded in 2016 to accelerate open source geospatial machine learning, SpaceNet is a nonprofit organization that runs data challenges and releases the training datasets, baseline algorithms, winning algorithms, and detailed evaluations under an open source license. They have organized six data challenges to date, each focusing on a different problem that applies machine learning to satellite imagery to solve complex mapping problems.

BigEarthNet is a new large-scale benchmark geospatial training data consisting of multi-label land cover classes in ten European countries. 

Radiant Earth Foundation, the leader in enabling access to geospatial training data, is pleased to announce the availability of the BigEarthNet large-scale benchmark archive through Radiant MLHub, the world’s first open library dedicated to Earth observation (EO) training data.

The BigEarthNet archive consists of 590, 326 Sentinel-2 image patches with spectral bands at 10, 20, and 60-meter resolution. The satellite images were acquired in different seasons between June 2017 and May 2018 over Austria, Belgium, Finland, Ireland, Kosovo, Lithuania, Luxembourg, Portugal, Serbia, and Switzerland.