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

TL;DR

This paper benchmarks deep learning methods for glacier calving front delineation in SAR images, highlighting current error margins and the need for improved techniques compared to human annotators.

Contribution

It provides a comparative analysis of deep learning systems versus human annotations for glacier front delineation in SAR imagery.

Findings

Deep learning systems have errors up to 221 m.

Human annotators deviate by only 38 m.

Further research is needed to improve automated delineation accuracy.

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.