Large-Scale Avalanche Mapping from SAR Images with Deep Learning-based Change Detection

2026-03-23

AI Summary

This study addresses the severe threats posed by rapid mass-movement hazards such as avalanches to lives, infrastructure, and ecosystems by proposing an end-to-end deep learning approach that formulates avalanche mapping as a single-modality, bi-temporal change detection task using Sentinel-1 synthetic aperture radar (SAR) imagery. It presents the first systematic validation of SAR-only-based avalanche identification, introduces an adjustable thresholding strategy to balance precision and recall, and releases a multi-region annotated dataset to establish a reproducible benchmark. Experimental results demonstrate an F1-score of 0.8061 under a conservative configuration and an F2-score of 0.8414 in a high-recall setting, with an avalanche polygon hit rate of 80.36%, significantly enhancing the detection capability for small-scale and marginal avalanches.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data

Abstract

Accurate change detection from satellite imagery is essential for monitoring rapid mass-movement hazards such as snow avalanches, which increasingly threaten human life, infrastructure, and ecosystems due to their rising frequency and intensity. This study presents a systematic investigation of large-scale avalanche mapping through bi-temporal change detection using Sentinel-1 synthetic aperture radar (SAR) imagery. Extensive experiments across multiple alpine ecoregions with manually validated avalanche inventories show that treating the task as a unimodal change detection problem, relying solely on pre- and post-event SAR images, achieves the most consistent performance. The proposed end-to-end pipeline achieves an F1-score of 0.8061 in a conservative (F1-optimized) configuration and attains an F2-score of 0.8414 with 80.36% avalanche-polygon hit rate under a less conservative, recall-oriented (F2-optimized) tuning. These results highlight the trade-off between precision and completeness and demonstrate how threshold adjustment can improve the detection of smaller or marginal avalanches. The release of the annotated multi-region dataset establishes a reproducible benchmark for SAR-based avalanche mapping.

Problem

• avalanche mapping
• change detection
• SAR imagery
• mass-movement hazards
• large-scale monitoring

Innovation

• SAR change detection
• deep learning
• avalanche mapping
• bi-temporal imaging
• benchmark dataset