Happy New Year! Highlights in December include the availability of geospatial models on Radiant MLHub, new land cover tools, tutorials, more data competitions, STAC integration in Leafmap and catalog conversion in Earth Engine, AMLD Africa 2021, Neur IPS 2021 recordings, and other impactful machine learning for earth observation resources.
Methodologies & Applications
- SyntEO: Synthetic dataset generation for Earth observation with deep learning – demonstrated for offshore wind farm detection
- Next Day Wildfire Spread: A machine learning data set to predict wildfire spreading from remote-sensing data
- Deep learning approach for Sentinel-1 surface water mapping leveraging Google Earth Engine
- Using WorldView-3 multispectral and shortwave infrared satellite imagery to understand urban deprivation through land cover in Nairobi, Kenya
- Automated delineation of agricultural field boundaries from Sentinel-2 images using recurrent residual U-Net
- Deep learning-based landslide susceptibility mapping
- Fast urban land cover mapping exploiting Sentinel-1 and Sentinel-2 data
- Mapping tree cover loss in Forestière Guinée using Landsat analysis ready data and a regionally calibrated, annual forest change detection model
- Applying interactive visualization and representation analysis to guide interpretation of glacier segmentation models
- Efficient spatio-temporal weather forecasting using U-Net
- Mapping standing dead trees in temperate montane forests using a pixel- and object-based image fusion method and stereo WorldView-3 imagery
- A multi-sensor approach for characterising human-made structures by estimating area, volume and population based on sentinel data and deep learning
- A survey of future prospects for deep learning of hyperspectral image classification
- A methodological guide to using machine learning in the spatial context
- Deep Learning for spatiotemporal modeling of urbanization
- Global maps of cropland extent and change show accelerated cropland expansion in the twenty-first century
Access & Validation
- ICIMOD: Future sea-level rise projections for tide gauge locations in South Asia
- Next Day Wildfire Spread: A dataset to predict wildfire spreading from remote-sensing data
- Global cropland expansion in the 21st century: Time-series at 30-m spatial resolution dataset
- Maxar Open Data Program releases data on Colorado wildfires and Indonesia’s Mount Semeru eruption for noncommercial use.
Standards Research & Innovation
- A belated introduction to stactools
- STAC Node Validator v1.2.0 released with additional updates in v1.2.1 and bug fixes in v1.2.2
- Towards a Cloud-Native Geospatial standards baseline
- STAC ML Model Extension
Tutorials, Webinars (recorded) & Resources
TUTORIALS
- Applying Raster Vision: A geospatial deep learning framework, to a semantic segmentation problem
- Stereo imagery, digital surface models, and machine learning for vegetation management
- Use deep learning to determine the extent of mangrove forests in Mumbai, India and see how their footprints have changed over time.
- How to use the new Azure machine Learning feature for object detection
- A step-by-step tutorial about learning how to perform feature reduction and then the classification of hyperspectral data using Support Vector Machines
- How to read, visualize and analyze digital elevation models in Python
- Getting started with geospatial analysis on SageMaker Studio Lab
- Accessing open ocean data: with Aqualink & pyaqua
- Dask live by Coiled tutorial by Naty Clementi
- Machine learning with Earth observation data: Case studies with semantic segmentation and regression
WEBINARS
- IIEPGW: Predicting and mapping multidimensional poverty measurement with Neeti Pokhriyal and Nattapong Puttanapong
- NeurIPS 2021: Deep learning for spatiotemporal modeling of urbanization with Tang Li (Best Paper Award)
- NASEM Earth and Life Studies: Accelerating the analysis of geographic change
- Google Developer Groups: Simulation City, transforming real-world data into a 3D Digital Twin of the Earth with Luke Barrington
- NASA Earth Data: Using geospatial data to evaluate climate hazards and inform environmental justice
- GeoHero: Machine learning bias in GIS applications with Caroline Gevaert
- AMLD Africa 2021 presentations playlist
- NeurIPS 2021 presentations playlist
PODCASTS
- AI in Africa: Machine learning for Earth observation with Hamed Alemohammad, Abba Barde, and Joyce Nabende
- Scene from Above: Inclusion in the Earth observation fields with Flávia Mendes
- Minds Behind Maps: Iceye SAR data and building a scalable product with Charles Blanchet
- Minds Behind Maps: Global Forest Watch, Design and Storytelling with Dan Hammer
- Scene from Above: Latest news, including geospatial model availability on Radiant MLHub, Earth Engine conversion to STAC catalog, and more
- Geomob: The politics of Geo with Ivan Sanchez
- Project Geospatial: Intro to machine learning with Chul Gwon
- Geomob: Looking back at the geospatial activities of 2021
RESOURCES
- Geospatial models now available in Radiant MLHub
- How deep learning and Earth observation promises support of Sustainable Development Goals
- Available Now: Machine learning for Earth observation online course
- Digital Earth Africa launches Fractional Cover, analysis service that describes landscape by providing a classification of groundcover
- NASA Harvest-GEOGLAM releases Earth observations indicators tool for in-season agricultural monitoring
- Local tile server for geospatial raster provides option to choose raster bands to visualize for comparisons, details from Bane Sullivan’s Twitter
- How deep generative modeling can help accelerate and scale simulation of weather patterns
- Accepted papers available for access from NeurIPS 2021 Workshop: Tackling Climate Change with Machine Learning
- IRCARI Global Top 100: 100 projects solving problems related to the 17 United Nations Sustainable Development Goals with the application of Artificial Intelligence
- Road map for research on responsible artificial intelligence for development (AI4D) in African countries: The case study of agriculture
- New satellite connectivity and geospatial capabilities with Azure Space
- International Water Management Institute uses Earth Engine for climate crisis and food security in South Asia
- Australian state launches machine learning tree mapping tools
- Call for manuscripts (due Feb 28) for Special Issue “Recent Advances for Crop Mapping and Monitoring Using Remote Sensing Data”
- Leafmap adds Inspector tool to get pixel values from STAC and COG images
Data Challenges
- AI4FoodSecurity Challenge, South Africa and Germany (ends Jan 9)
- DrivenData: On Cloud N: Cloud Cover Detection Challenge (ends Feb 7)
- WiDS 2022 Datathon (ends Feb)
- Omdena – Using AI to Identify Optimal Locations Spots for Floating Solar Installations (ends Feb)
- DrivenData: Snowcast Showdown, Development Stage (ends March 15)
Conferences & Workshops
- IARAI Workshop on Complex Data Challenges in Earth Observation (Jan 11)
- ESIPFed January meeting (Jan 18-21)
- 21st Conference on Artificial Intelligence for Environmental Science (Jan 23-27)
- Machine Learning and Artificial Intelligence to Advance Earth System Science: Opportunities and Challenges (Feb 7-11)
- Women in Data Science Worldwide Conference (Mar 7)
- ECMWF Machine Learning Workshop (Mar 29- April 1)
- North51: RECOMBINATION (May 4-6)
What are we missing? Contact Louisa@radiant.earth