Snow avalanche segmentation in SAR images with Fully Convolutional Neural Networks

TL;DR

This paper presents a deep learning approach using Fully Convolutional Neural Networks to detect snow avalanches in Sentinel-1 SAR images, significantly outperforming traditional radar signal processing methods.

Contribution

The authors develop and train a neural network model that improves avalanche detection accuracy in SAR images, reducing the gap between automated detection and human expert performance.

Findings

F1 score of over 66% on test data

Outperforms existing radar-based detection algorithms with 38% F1 score

Detects small avalanches and finds previously unlabelled ones

Abstract

Knowledge about frequency and location of snow avalanche activity is essential for forecasting and mapping of snow avalanche hazard. Traditional field monitoring of avalanche activity has limitations, especially when surveying large and remote areas. In recent years, avalanche detection in Sentinel-1 radar satellite imagery has been developed to improve monitoring. However, the current state-of-the-art detection algorithms, based on radar signal processing techniques, are still much less accurate than human experts. To reduce this gap, we propose a deep learning architecture for detecting avalanches in Sentinel-1 radar images. We trained a neural network on 6,345 manually labelled avalanches from 117 Sentinel-1 images, each one consisting of six channels that include backscatter and topographical information. Then, we tested our trained model on a new SAR image. Comparing to the manual labelling (the gold standard), we achieved an F1 score above 66\%, while the state-of-the-art detection algorithm sits at an F1 score of only 38\%. A visual inspection of the results generated by our deep learning model shows that only small avalanches are undetected, while some avalanches that were originally not labelled by the human expert are discovered.