Explore the call to publish your training data on Radiant MLHub for NeurIPS 2021 new datasets and benchmark track, a range of datasets available on Radiant MLHub and other platforms, data challenges, podcasts showcasing leading individuals in the ML4EO sector, and more highlights this April.
Methodologies & Applications
- Applying machine learning methods to detect convection using satellite data for precipitation forecasts
- A machine learning approach to human footprint index estimation for sustainable development
- Comparing machine learning algorithms to evaluate Sentinel-2 and Landsat 8 data for land cover/use and classification
- Mapping spatial distribution and biomass of algal blooms across coastal areas in Ireland using machine learning and satellite imagery
- Earthquake-induced landslide susceptibility mapping with a one-class-classifier based in southwest China
- Model trained on satellite imagery developed to detect pollution hotspots at 300 m resolution
- Convolutional neural networks and WorldView-3 satellite images used to map Brazil nut trees
- Building footprint segmentation through a deep learning framework using SpaceNet 6
- Machine-learned 3D building vectorization from WorldView imagery
- Calculating distributional uncertainty of deep learning models for remote sensing
- Crop-type segmentation using contrastive learning and Sentinel-2 satellite data
- Producing physically-consistent imagery using pix2pixHD GAN for coastal visualization
- Challenges and opportunities in precision irrigation decision-support systems with satellite data and statistical/machine learning models
- Mapping the invasive plant species Opuntia stricta in arid environments of Kenya with Sentinel-2 imagery and machine learning classifiers
- Examining aboveground biomass across boreal northwestern North America using satellite data and models
- A meta-analysis of Convolutional Neural Networks for remote sensing applications
Data Access & Validation
- Four new datasets available on Radiant MLHub: SEN12-Flood, a crop-type dataset for Central Asia, smallholder cashew plantations in Benin, and crop-type classification in Rwanda
- EarthNet2021: A dataset for Earth surface and localized impact forecasting
- A high-resolution daily gridded meteorological dataset for Serbia made by Random Forest Spatial Interpolation
- USGS Land Change Monitoring, Assessment, and Project initiative releases the first update of its Conterminous United States science products, adding two years of data
- Digital Earth Africa datasets now listed on AWS open data registry
- High resolution population density maps and demographic estimates by CIESIN and Facebook available on AWS
- OpenTopography’s catalog of global topographic data now available in COG
- A multiple species, million-phenotype agronomic plant dataset for Australia by researchers at Australian National University
- A new dataset for global irrigation areas from 2001 to 2015 is developed by University of Berkley, California and Polytechnic University of Milan researchers
Standards Research & Innovation
- STAC-Validator 2.0.0 released for python to validate STAC json files against the STAC spec
- Use StackSTAC to turn a a STAC catalog into a dask-based xaraay
Tutorials, Webinars (recorded) & Resources
TUTORIALS
- Generating map tiles at different zoom levels using Gdal2tiles in python to address high-resolution raster data concerns
- How to apply Laguerre-Vornoi tessellation to Demographic and Health Survey (DHS) data for analysis geospatial analysis and machine learning tasks
- A beginners guide to labelling SAR for machine learning from Azavea
- How to leverage geolocation data for machine learning through feature engineering and visualization
- How to access iSDAsoil property and nutrient maps for Africa, Sentinel-2 cloud-free bands and terrain variables, and how to compute with it efficiently
- Using Google Earth Engine and Google Cloud AutoML to predict malaria vector populations
- Using satellite imagery to train a CNN to classify ships
- Cloud-based processing of satellite images using STAC, COGs and data cubes in R
- How to use GEOGLAM Global Crop Monitor and ways to use the data, part 1
WEBINARS
- Coiled’s webinar with Chelle Gentemann and Hugo Bowne-Anderson on distributed data science and oceanography with Dask
- AI for Good webinar on geospatial artificial intelligence/machine learning applications and policies from a global perspective
- AI for Good workshop with Polytechnic University of Milan on satellite data analysis and machine learning classification with QGIS, part 1
- Stanford HAI’s Stefano Ermon on artificial intelligence for sustainable development
- Climate Change AI webinar on the tackling climate change with machine learning ICML 2021 workshop
- Geopython Conference 2021 workshop on Google Earth Engine and geemap with Quisheng Wu
PODCASTS
- Scene From Above with Vincent Sarago: COGs, Remote Pixel and Development Seed
- Minds Behind Maps with Ian Schuler: Solving difficult problems with global partners
- MapScaping with Nadine Alameh: What open geospatial standards are
- MapScaping with Markus Müller: Using super-resolution for Earth observation modeling
- Degrees of Freedom with Bruno Sánchez-Andrade Nuño: Struggles in finding the right path for impactful work
- GeoMob Podcast with Aleks Buczkowski and Muthu Kumar: Geoawesomeness turns ten
- GeoMob Podcast with Eric Rodenbeck and Alan McConchie: The evolution of Stamen
RESOURCES
- Call to publish your training data on Radiant MLHub for NeurIPS 2021 new datasets and benchmark track that requires data documentation and an open repository
- Microsoft launched the Planetary Computer to scale environmental sustainability work with the power of the cloud, currently in private beta but anyone can use the data and STAC API
- Geo-Microsoft Planetary Computer Programme invites the GEO community to be early adopters
- Using cloud-native repositories for big scientific data to overcome infrastructure challenges
- Africa GeoPortal releases Nigeria GeoPortal, featuring various dashboards
- Igor Ivanov: Harnessing machine learning skills to reduce damages from tropical storms. Q&A with the First Place winner of the Radiant Earth Tropical Cyclone Wind Estimation Data Competition
- Online satellite imagery tool showcases three decades of coastal change in Australia using Digital Earth Australia’s Coastlines dataset
- Enabling Crop Analytics At Scale (ECAAS) advances crop analytics that integrates remote sensing and ML, requests proposals
- New organization, PLACE, to map data in the public interest through a sustainability model
- UNEP and Google partner to hunt for plastic pollution with machine learning in the Mekong rivers
- A collection of remote sensing collections that are free to use by Sentinel Hub
- New project to use artificial intelligence and machine learning to detect and monitor permafrost thawing
- The awesome-gee-community-datasets catalog now has a webpage to ease browsing and searching
Data Challenges
- Earth Vision 2021 Challenges: FloodNet Semantic Segmentation (ends May 15) & DynamicEarthNet Land Cover Change Detection (ends June 1)
- Commonwealth Secretariat and Satellite Applications Catapult – Hack the Planet (ends May 31)
- Vision for Agriculture- Agriculture-Vision Prize Challenge (ends June 5)
- Lacuna – Correct Field Detection Challenge (ends June 13)
- NASA IMPACT and IEEE GRSS- ETCI 2021 Flood Detection Challenge (ends June 30)
- 2021 OSGeo UN Committee Educational Challenges – Training on Satellite Data Analysis and Machine Learning with QGIS & Workshop Material for pgRouting (ends June 14)
- ESA- Digital Twin Earth Challenge (ends July 19)
- Copernicus Masters Competition (ends July 19)
- EarthNet 2021: A machine learning challenge for Earth surface and localized impact forecasting
- The 2020 Climate Informatics Hackathon – Public Challenge (open)
Conferences & Workshops
- AGU Ocean Visions Summit 2021(May 18-21)
- First International Geoscience and Remote Sensing Symposium of GRSS-Chile (June 2)
- CVPR 2021 (June 19-25)
- IGARSS’ 2021 Symposium (July 12-16)
- 2021 FOSS4G (September 27 – October 2)
- ICML 2021 Workshop on Tackling Climate Change with Machine Learning (July 23 or 24 TBD)
- Data-driven Humanitarian Mapping Workshop, KDD 2021 (Aug 14-18)
What are we missing? Contact Louisa@radiant.earth