Multi-temporal Calving Front Segmentation

TL;DR

This paper introduces a multi-temporal deep learning approach for calving front segmentation in satellite imagery, improving accuracy by leveraging temporal information across image sequences.

Contribution

It proposes a novel multi-temporal processing method integrated into Tyrion architecture, achieving state-of-the-art performance on calving front segmentation benchmarks.

Findings

Achieved a Mean Distance Error of 184.4 m

Attained a mean Intersection over Union of 83.6

Outperformed previous models on the CaFFe dataset

Abstract

The calving fronts of marine-terminating glaciers undergo constant changes. These changes significantly affect the glacier's mass and dynamics, demanding continuous monitoring. To address this need, deep learning models were developed that can automatically delineate the calving front in Synthetic Aperture Radar imagery. However, these models often struggle to correctly classify areas affected by seasonal conditions such as ice melange or snow-covered surfaces. To address this issue, we propose to process multiple frames from a satellite image time series of the same glacier in parallel and exchange temporal information between the corresponding feature maps to stabilize each prediction. We integrate our approach into the current state-of-the-art architecture Tyrion and accomplish a new state-of-the-art performance on the CaFFe benchmark dataset. In particular, we achieve a Mean Distance Error of 184.4 m and a mean Intersection over Union of 83.6.