Comparison Study: Glacier Calving Front Delineation in Synthetic Aperture Radar Images With Deep Learning

Research output: Contribution to journal › Research article › Contributed › peer-review

Contributors

  • Nora Gourmelon - (Author)
  • Konrad Heidler - (Author)
  • Erik Loebel - , Chair of Geodetic Earth System Research (Author)
  • Daniel Cheng - (Author)
  • Julian Klink - (Author)
  • Anda Dong - (Author)
  • Fei Wu - (Author)
  • Noah Maul - (Author)
  • Moritz Koch - (Author)
  • Marcel Dreier - (Author)
  • Dakota Pyles - (Author)
  • Thorsten Seehaus - (Author)
  • Matthias Braun - (Author)
  • Andreas Maier - (Author)
  • Vincent Christlein - (Author)

Abstract

Continuous monitoring of glacier calving fronts is essential for sea level rise projections. This study benchmarks Deep Learning systems for front delineation in Synthetic Aperture Radar imagery. While Deep Learning systems exhibit errors up to 221 m, human annotators deviate by only 38 m, underscoring the need for further research.

Details

Original languageEnglish
JournalIEEE Transactions on Pattern Analysis and Machine Intelligence
Publication statusE-pub ahead of print - 20 Apr 2026
Peer-reviewedYes

External IDs

ORCID /0000-0001-9874-9295/work/212490969

Keywords