ICEBERG: Imagery Cyberinfrastructure and Extensible Building-Blocks to Enhance Research in the Geosciences
NSF EarthCube: Collaborative Proposal
Award Number:1740581
Polar geosciences stands at the precipice of a revolution, one enabled by the confluence of cutting edge analytical tools, petabytes of high-resolution imagery, and an ever growing array of high performance computing resources. With these tools at hand, we can look beyond incremental improvements in our understanding of the polar regions; instead, fully capitalizing on these opportunities will radically change our capacity to address priority questions for NSF Geosciences. Near-real time datasets of geological and biological importance at the continental scale are within our reach if we create those critical cyberinfrastructure (CI) components that allow the geosciences community to exploit existing assets and establish a common workflow for reproducible imagery-enabled science. The research objective of this proposal is to understand the biological, geological, and hydrological functioning of the polar regions at spatial scales heretofore beyond the reach of individual PIs, and to develop tools for imagery-enabled science that can be applied globally. The resulting cyberinfrastructure, which we call ICEBERG — Imagery Cyberinfrastructure and Extensible Building-Blocks to Enhance Research in the Geosciences, is an extensible system for coupling open-source image analysis tools with the use of high performance and distributed computing (HPDC) for imagery-enabled geoscience research. Moving from megabytes to petabytes of available imagery requires new computational approaches to interpretation coupled with efficient use of HPDC resources. As the spatial resolution of satellite imagery has continued to shrink, we’ve seen the convergence of traditional pixel-based remote sensing and computer vision approaches based on object characteristics such as context and texture. Additionally, while the sheer volume of imagery poses challenges for processing, we now have enough training data to develop deep learning algorithms for imagery interpretation that can accelerate the speed of classification and geoscience research. To address these challenges, we propose an Integration project to (1) develop open source image classification tools tailored to high-resolution satellite imagery of the Arctic and Antarctic to be used on HPDC resources, (2) create easy-to-use interfaces to facilitate the development and testing of algorithms for application specific geoscience requirements, (3) apply these tools through four use cases that span the biological, hydrological, and geoscience needs of the polar community, (4) transfer these tools to the larger (non-polar) EarthCube community for continued community-driven development.