Predicting Natural Hazards with Neuronal Networks.

RSS Source
Matthias Rauter, Daniel Winkler

Gravitational mass flows, such as avalanches, debris flows and rockfalls arecommon events in alpine regions with high impact on transport routes. Withinthe last few decades, hazard zone maps have been developed to systematicallyapproach this threat. These maps mark vulnerable zones in habitable areas toallow effective planning of hazard mitigation measures and development ofsettlements. Hazard zone maps have shown to be an effective tool to reducefatalities during extreme events. They are created in a complex process, basedon experience, empirical models, physical simulations and historical data. Thegeneration of such maps is therefore expensive and limited to cruciallyimportant regions, e.g. permanently inhabited areas.

In this work we interpret the task of hazard zone mapping as a classificationproblem. Every point in a specific area has to be classified according to itsvulnerability. On a regional scale this leads to a segmentation problem, wherethe total area has to be divided in the respective hazard zones. The recentdevelopments in artificial intelligence, namely convolutional neuronalnetworks, have led to major improvement in a very similar task, imageclassification and semantic segmentation, i.e. computer vision. We use aconvolutional neuronal network to identify terrain formations with thepotential for catastrophic snow avalanches and label points in their reach asvulnerable. Repeating this procedure for all points allows us to generate anartificial hazard zone map. We demonstrate that the approach is feasible andpromising based on the hazard zone map of the Tirolean Oberland. However, moretraining data and further improvement of the method is required before suchtechniques can be applied reliably.

Stay in the loop.

Subscribe to our newsletter for a weekly update on the latest podcast, news, events, and jobs postings.