Home // GEOProcessing 2026, The Eighteenth International Conference on Advanced Geographic Information Systems, Applications, and Services // View article

Physics-Informed Signed Distance Fields for Flood Arrival Time Prediction and Evacuation Routing

Authors:
Yameng Guo
Seppe vanden Broucke

Keywords: Signed distance function; Flood prediction; Evacuation routing.

Abstract:
Accurate flood prediction with explicit arrival times is critical for effective evacuation planning. Traditional physics-based models provide reliable predictions but are computationally expensive, while existing data-driven approaches typically predict binary flood extent or water depth without explicit timing information. We propose a novel framework that represents flood evolution as a temporal Signed Distance Field (SDF), enabling efficient prediction of both flood boundaries and arrival times. Our approach combines spatial-temporal deep learning with physics-informed constraints in a variational optimization framework. The predicted SDF naturally supports time-dependent evacuation routing by providing continuous distance-to-flood information and explicit arrival time fields. The proposed idea will be compared with grid-based deep learning approaches as well as physical constraints approaches regarding real-time evacuation planning and arrival time accuracy.

Pages: 1 to 5

Copyright: Copyright (c) IARIA, 2026

Publication date: May 24, 2026

Published in: conference

ISSN: 2308-393X

ISBN: 978-1-68558-384-2

Location: Venice, Italy

Dates: from May 24, 2026 to May 28, 2026