Adaptive fine-tuning for transferring a U-net hydrography extraction model using K-means

Larry Stanislawski; Ethan J. Shavers; Neal J. Pastick; Philip T. Thiem; Shaowen Wang; Nattapon Jaroenchai; Zhe Jiang; Barry J. Kronenfeld; Barbara P. Buttenfield; Adam Camerer

Adaptive fine-tuning for transferring a U-net hydrography extraction model using K-means

By: Larry Stanislawski, Ethan J. Shavers, Neal J. Pastick, Philip T. Thiem, Shaowen Wang, Nattapon Jaroenchai, Zhe Jiang, Barry J. Kronenfeld, Barbara P. Buttenfield, and Adam Camerer

Metrics

Web analytics dashboard Metrics definitions

Links

More information: Publisher Index Page
Download citation as: RIS | Dublin Core

Abstract

The United States Geological Survey (USGS) coordinates the collection of hydrographic features derived from remotely sensed interferometric synthetic aperture radar (IfSAR) elevation and intensity data in Alaska. Hydrographic features are cartographic representations of surface water features such as stream, rivers, lakes, ponds, canals, etc. Collection and validation procedures involve complex automated and manual techniques that furnish snapshots of hydrographic vector data that exist during the IfSAR surveys. The dynamic nature of fluvial conditions warrants monitoring and updating hydrographic data, but extraction procedures for updates can be cost prohibitive. This paper overviews progress on automated workflows to extract hydrography from IfSAR data using deep learning methods trained and tested with USGS collected hydrography data. This research tests transfer learning methods on a well-performing U-net model trained on a 4600-square kilometer (sq km) base model area in northcentral Alaska. The base model is transferred and fine-tuned to regions in the target domain covering roughly 127,000 sq km. The target domain is subdivided into areas with similar hydrogeomorphic conditions using principal components and k-means clustering, and the base model is adaptively fine-tuned to each hydrogeomorphic class by selecting training watersheds from each cluster within the target domain. Results are compared with transfer learning that is fine-tuned with a random sample of watersheds in the target domain.

Suggested Citation

Stanislawski, L., Shavers, E.J., Pastick, N.J., Thiem, P.T., Wang, S., Jaroenchai, N., Jiang, Z., Kronenfeld, B.J., Buttenfield, B.P., and Camerer, A., 2024, Adaptive fine-tuning for transferring a U-net hydrography extraction model using K-means, Cartography and Geographic Information Society (CaGIS) and the University Consortium for Geographic Information Science (UCGIS) 2024 Symposium, Columbus, OH, June 3-6, 2024, 45, 6 p.

Additional publication details
Publication type	Conference Paper
Publication Subtype	Conference Paper
Title	Adaptive fine-tuning for transferring a U-net hydrography extraction model using K-means
Year Published	2024
Language	English
Publisher	The Cartography and Geographic Information Society (CaGIS) and the University Consortium for Geographic Information Science (UCGIS)
Contributing office(s)	Center for Geospatial Information Science (CEGIS)
Description	45, 6 p.
Conference Title	Cartography and Geographic Information Society (CaGIS) and the University Consortium for Geographic Information Science (UCGIS) 2024 Symposium
Conference Location	Columbus, OH
Conference Date	June 3-6, 2024