Automated Detection of Grounding Lines with Deep Learning and Phase Gradient Analysis
Friday, February 7, 2025
10:00 am - 11:00 am
The grounding line, where a glacier transitions from grounded ice to floating ice, is crucial for tracking sea level changes and glacier retreat. Traditional mapping of grounding lines relies on manual expert analysis, which is increasingly impractical given the growing volume of Antarctic glacier data. This study explores automating the process through a deep learning approach and a phase gradient-based technique. Using a dataset from five Antarctic glaciers (2020–2022), the research evaluates the strengths and limitations of these methods, compares them to manual mapping, and provides recommendations for optimal techniques. The study also examines long-term retreat trends and daily tidal motion-induced variability of the glaciers.
About the Speaker
Natalya Ross is a Ph.D. candidate in the Civil and Environmental Engineering Department at the University of Houston. As a research assistant at the UH National Center for Airborne Laser Mapping, she is involved in cryosphere research, focusing on modeling and monitoring Antarctic conditions. Her work leverages advanced radar techniques, deep learning methodologies, and complex physical and numerical simulations.
- Location
- Online
- Cost
- Free