Item: DEEP LEARNING FOR REAL-TIME AVALANCHE DETECTION IN WEBCAM IMAGES
-
-
Title: DEEP LEARNING FOR REAL-TIME AVALANCHE DETECTION IN WEBCAM IMAGES
Proceedings: International Snow Science Workshop Proceedings 2023, Bend, Oregon
Authors:
- James Fox [ Department of Computer Science, University of Innsbruck, Austria ]
- Anna Siebenbrunner [ Lo.La Peak Solutions GmbH, Innsbruck, Austria ] [ Faculty of Geo- and Atmospheric Sciences, University of Innsbruck, Austria ]
- Sandra Reitinger [ Department of Computer Science, University of Innsbruck, Austria ]
- David Peer [ Department of Computer Science, University of Innsbruck, Austria ] [ DeepOpinion, Innsbruck, Austria ]
- Antonio Rodríguez-Sánchez [ Department of Computer Science, University of Innsbruck, Austria ]
Date: 2023-10-08
Abstract: Continuous snow avalanche monitoring is essential to enable rapid responses to avalanche incidents. However, existing satellite-based methods of remote avalanche detection are typically unsuitable for ongoing monitoring due to long satellite revisit intervals and insufficient spatial resolution. This paper proposes a novel approach to automating avalanche detection via analysis of webcam streams with deep learning models. To assess the viability of this approach, we trained convolutional neural networks on a publicly-released dataset of 4090 mountain photographs and achieved avalanche detection F1 scores of 92.9% per image and 64.0% per avalanche. Notably, our models do not require a digital elevation model, enabling straightforward integration with existing webcams in new geographic regions. The paper concludes with findings from an initial case study conducted in the Austrian Alps and our vision for operational applications of trained models. The code and dataset are available at https://github.com/j-f-ox/avalanche-detection
Object ID: ISSW2023_P3.47.pdf
Language of Article: English
Presenter(s): Anna Siebenbrunner
Keywords: avalanche detection, avalanche monitoring, avalanche, snow avalanche, remote sensing, deep learning
Page Number(s): 1504 - 1511
-