FloodNet: A Multilevel Multimodal Fusion Network With Semantic Consistency Constraint Strategy for Flood Segmentation

Abstract

Flood segmentation using synthetic aperture radar (SAR) images is essential for determining the extent of inundation areas, and informing subsequent management recommendations. However, existing networks for flood segmentation using single modality SAR images often face inherent challenges, including interference from terrain shadows and water-like surfaces, leading to degraded segmentation performance. In this study, we introduced a multilevel multimodal fusion network (FloodNet), in which an adaptive gated feature fusion module (AGFFM) is designed to integrate multimodal features from Sentinel-1 SAR images, digital elevation model (DEM), and Joint Research Center Global Surface Water (JRC-gsw). Furthermore, we proposed a semantic consistency constraint strategy to alleviate the blurring of water edges during the prediction process. Experiments on two publicly available flood datasets, C2S-Flood and ETCI-Flood, demonstrate the competitive performance of the proposed FloodNet compared with other state-of-the-art single- and multimodal networks.