WatNet Surface Water Mapping

WatNet applies convolutional neural networks to Sentinel-2 scenes to delineate surface water at high spatial detail (10 m resolution). The current preview focuses on Berlin, highlighting how consistent pixel-wise classification can accelerate localized water monitoring efforts.

WatNet: a deep learning application for surface water features on Sentinel-2 satellite imagery. Frames cycle slowly to highlight temporal consistency along the Spree on 2023-06-05 10:16:01.

Reference: Luo, X., Tong, X., & Hu, Z. (2021). An applicable and automatic method for earth surface water mapping based on multispectral images. International Journal of Applied Earth Observation and Geoinformation, 103, 102472. https://doi.org/10.1016/j.jag.2021.102472