H34B:
Applications of Machine Learning and Data Cubes in Remote Sensing II


Submit an Abstract to this Session


Session ID#: 62355

Session Description:
Rich satellite observations over large spatio-temporal scales provide invaluable big datasets for monitoring the Earth and its changing environment. Advancements in data-driven techniques, in particular machine learning (ML), enables us to harness these datasets and advance geoscience and environmental sciences in unique and unprecedented ways. Such applications include scientific discoveries that lead to new findings and practical applications that lead to improved decision making. Furthermore, development of new data hosting technologies such as Data Cubes makes it easier to apply ML and time series analysis at larger spatial scales on the cloud computing platforms or supercomputers. This session solicits recent advancements of using ML and remote sensing for applications in geoscience and environmental sciences and opportunities that data cubes provide. We welcome presentations in either technology advancements or application-driven case studies in the geoscience context. We also welcome presentations that address challenges in generating representative training data for ML algorithms.
Primary Convener:  Hamed Alemohammad, Radiant Earth Foundation, Washington, United States
Conveners:  Steven Brumby, Descartes Labs, Inc., Los Alamos, United States; World Resources Institute, Washington, DC, United States, Brian Killough, NASA Langley Research Center, Hampton, VA, United States and Kaiyu Guan, University of Illinois Urbana-Champaign, Agroecosystem Sustainability Center, Institute for Sustainability, Energy, and Environment, Urbana, United States
Primary Liaison:  Hamed Alemohammad, Radiant Earth Foundation, Washington, United States
Chairs:  Hamed Alemohammad, Radiant Earth Foundation, Washington, United States, Brian Killough, NASA Langley Research Center, Hampton, VA, United States and Steven Brumby, Descartes Labs, Inc., Los Alamos, United States
OSPA Liaison:  Hamed Alemohammad, Radiant Earth Foundation, Washington, United States
Co-Organized with:
Hydrology, and Earth and Space Science Informatics

Cross-Listed:
  • A - Atmospheric Sciences
  • B - Biogeosciences
  • GC - Global Environmental Change
  • IN - Earth and Space Science Informatics

Proposed Co-Organized Session with:
  • B - Biogeosciences
  • GC - Global Environmental Change
  • GH - GeoHealth
  • IN - Earth and Space Science Informatics
Index Terms:

0434 Data sets [BIOGEOSCIENCES]
1855 Remote sensing [HYDROLOGY]
1926 Geospatial [INFORMATICS]
1942 Machine learning [INFORMATICS]

Abstracts Submitted to this Session:

Fully Convolutional Networks for Landcover Classification and Landcover Change (381919)
Brookie P Guzder-Williams, World Resources Institute, San Francisco, CA, United States
ISOSCELES: Data Driven Sampling for Efficient Convolutional Neural Network Training (447342)
Benjamin Swan, Oak Ridge National Laboratory, Geospatial Science and Human Security Division, Oak Ridge, United States, Melanie Laverdiere, Oak Ridge National Laboratory, Oak Ridge, TN, United States and Lexie Yang, Oak Ridge National Laboratory, Geospatial Science and Human Security, Oak Ridge, TN, United States
Random Forest Classification of Regions in Colombia and Uruguay using Open Data Cube (ODC) Framework (368241)
Syed R Rizvi, Analytical Mechanics and Associates, Inc., Norfolk, VA, United States and Brian Killough, NASA Langley Research Center, Hampton, VA, United States
High-resolution downscaling of multi-band satellite images without co-located high-resolution data: a new approach based on training images (463701)
Grégoire Mariethoz1, Fabio Oriani1 and Matthew McCabe2, (1)University of Lausanne, Faculty of Geosciences and Environment, Institute of Earth Surface Dynamics, Lausanne, Switzerland, (2)King Abdullah University of Science and Technology, Hydrology, Agriculture and Land Observation (HALO) Laboratory, Division of Biological and Environmental Sciences and Engineering, Thuwal, Saudi Arabia
Convolutions of random patches as a generalizable featurization for multi-domain prediction using remote sensing imagery (396485)
Ian Wesley Bolliger1, Tamma Carleton2,3, Solomon M Hsiang4, Jonathan Kadish5, Jonathan Proctor6, Ben Recht7, Esther Rolf8 and Vaishaal Shankar7, (1)University of California Berkeley, Energy and Resources Group, Berkeley, CA, United States, (2)University of California, Berkeley, Berkeley, CA, United States, (3)University of Chicago, Economics, Chicago, IL, United States, (4)University of California, Berkeley, Goldman School of Public Policy, Berkeley, United States, (5)Elastic Data Lab LLC, Portland, OR, United States, (6)Harvard University, Center for the Environment, Cambridge, United States, (7)University of California Berkeley, Electrical Engineering and Computer Science, Berkeley, CA, United States, (8)University of California Berkeley, Electrical Engineering and Computer Science, Berkeley, United States
NeMO-Net – The Neural Multi-Modal Observation & Training Network for Global Coral Reef Assessment (410894)
Alan Sheng Xi Li1, Ved Chirayath2, Kamalika Das3, Michal Segal-Rozenhaimer2, Jarrett van den Bergh1 and Juan Torres4, (1)NASA Ames Research Center, Moffett Field, CA, United States, (2)NASA Ames Research Center, Moffett Field, United States, (3)University of Maryland Baltimore County, Baltimore, MD, United States, (4)ARC-SGE)[Bay Area Environmental Research Institute], Mountain View, CA, United States
A Machine Learning Approach to Estimate Daily Air Temperature Globally (441675)
Roozbeh Raoufi, Northeastern University, Department of Civil and Environmental Engineering, Boston, MA, United States, Edward Beighley, Northeastern University, Civil and Environmental Engineering, Boston, United States and Amy V Mueller, Massachusetts Institute of Technology, Cambridge, MA, United States

See more of: Hydrology