Explainable Flood Segmentation on Sentinel-1 SAR Imagery: A Comparative Study of CNN and Transformer Architectures

• URL: http://arxiv.org/abs/2606.16302v1
• Date: Mon, 15 Jun 2026 07:06:30 GMT
• Title: Explainable Flood Segmentation on Sentinel-1 SAR Imagery: A Comparative Study of CNN and Transformer Architectures
• Abstract summary: This study provides a comprehensive comparison of convolutional neural network (CNN) and vision transformer architectures for multi-class flood segmentation.<n>Three state-of-the-art (SOTA)CNN-based models, U-Net, U-Net++, and DeepLabV3 with ResNet-34 backbone, and three SegFormer variants (b0,b1,b2) were evaluated.<n>The results demonstrate that SegFormer-b2 significantly outperforms the U-Net baseline on the ETCI dataset.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Rapid and accurate flood prediction is essential for disaster response and mitigation planning. Synthetic Aperture Radar (SAR) sensors in satellites are well-suited for this purpose because they operate independently of weather and daylight conditions. Although SAR-based data enable all-weather flood monitoring, distinguishing flooded land from permanent water remains a significant challenge, particularly when flooding is defined strictly as inundated land. This study provides a comprehensive comparison of convolutional neural network (CNN) and vision transformer architectures for multi-class flood segmentation using Sentinel-1 SAR imagery, specifically trained to separate flooded land from permanent water bodies and land. Three state-of-the-art (SOTA)CNN-based models, U-Net, U-Net++, and DeepLabV3 with ResNet-34 backbone, and three SegFormer variants (b0,b1,b2) were evaluated in two benchmark datasets, the ETCI NASA dataset and SenFloods11, using scene-based data splits to ensure a realistic assessment of spatial generalization. The results demonstrate that SegFormer-b2 significantly outperforms the U-Net baseline on the ETCI dataset (higher flood IoU across all 7 test scenes in the Wilcoxon signed-rank test), while after fine-tuning on Sen1Floods11, the advantage narrows to within the range of scene variability and is concentrated in spatially fragmented flood events. The study includes both qualitative and quantitative explainability techniques to visually comprehend model decisions and systematically assess prediction reliability. Qualitative analysis reveals that SegFormer-b2 produces more spatially coherent Grad-CAM activations focused on flood-relevant features, while U-Net generates more informative uncertainty estimates along flood boundaries.

Related papers

• Large scale cross-regional remote sensing flood monitoring framework for operative mapping and impact analysis [16.57552889738507]
  This study presents an end-to-end multimodal framework for sustainable flood monitoring and damage assessment in Russian Federation territories.<n>Two strategies for water surface detection under limited data conditions were compared: a supervised U-Net++ model and the self-supervised AnySat architecture.<n>The best flood area predictions were used to estimate flood impact in urban areas in terms of the area affected, material damage, casualties, ecological and agricultural impact.
  arXiv  Detail & Related papers  (2026-07-30T15:50:23Z)
• Topology-Informed Neural Networks for Flood Detection in Optical and Synthetic Aperture Radar Imagery [10.268453073867622]
  Rapid and accurate flood detection is crucial for emergency response and timely mitigation of human and economic loss.<n>Many flood events remain challenging to detect because cloud cover obscures optical satellite imagery.<n>We show that topological descriptors carry meaningful flood signals independently and complement existing networks to yield more robust and interpretable flood detection systems.
  arXiv  Detail & Related papers  (2026-06-24T16:52:28Z)
• Towards Realistic Open-Vocabulary Remote Sensing Segmentation: Benchmark and Baseline [52.65099689153431]
  Open-vocabulary remote sensing image segmentation (OVRSIS) remains underexplored due to fragmented datasets, limited training diversity, and the lack of evaluation benchmarks.<n>We propose textitOVRSISBenchV2, a large-scale and application-oriented benchmark for OVRSIS.<n>Our results highlight the importance of realistic benchmark design and the effectiveness of perturbation-based transfer for OVRSIS.
  arXiv  Detail & Related papers  (2026-04-17T02:49:46Z)
• WUTDet: A 100K-Scale Ship Detection Dataset and Benchmarks with Dense Small Objects [68.44670476527155]
  WUTDet is a large-scale ship detection dataset with 100,576 images and 381,378 annotated ship instances.<n>We systematically evaluate 20 baseline models from three mainstream detection architectures, namely CNN, Transformer, and Mamba.<n>Results on Ship-GEN show that models trained on WUTDet exhibit stronger generalization under different data distributions.
  arXiv  Detail & Related papers  (2026-04-09T03:28:05Z)
• Efficient Self-Supervised Learning for Earth Observation via Dynamic Dataset Curation [67.23953699167274]
  Self-supervised learning (SSL) has enabled the development of vision foundation models for Earth Observation (EO)<n>In EO, this challenge is amplified by the redundancy and heavy-tailed distributions common in satellite imagery.<n>We propose a dynamic dataset pruning strategy designed to improve SSL pre-training by maximizing dataset diversity and balance.
  arXiv  Detail & Related papers  (2025-04-09T15:13:26Z)
• Deep Self-Supervised Disturbance Mapping with the OPERA Sentinel-1 Radiometric Terrain Corrected SAR Backscatter Product [41.94295877935867]
  Mapping land surface disturbances supports disaster response, resource and ecosystem management, and climate adaptation efforts.<n> Synthetic aperture radar (SAR) is an invaluable tool for disturbance mapping, providing consistent time-series images of the ground regardless of weather or illumination conditions.<n>NASA's Observational Products for End-Users from Remote Sensing Analysis (OPERA) project released the near-global Radiometric Terrain Corrected SAR backscatter from Sentinel-1 (RTC-S1) dataset in October 2023.<n>In this work, we utilize this new dataset to systematically analyze land surface disturbances.
  arXiv  Detail & Related papers  (2025-01-15T20:24:18Z)
• Using Multi-Temporal Sentinel-1 and Sentinel-2 data for water bodies mapping [40.996860106131244]
  Climate change is intensifying extreme weather events, causing both water scarcity and severe rainfall unpredictability. This paper aims to provide valuable insights for comprehensive water resource monitoring under diverse meteorological conditions.
  arXiv  Detail & Related papers  (2024-01-05T18:11:08Z)
• Leveraging Citizen Science for Flood Extent Detection using Machine Learning Benchmark Dataset [0.9029386959445269]
  We create a labeled known water body extent and flooded area extents during known flooding events covering about 36,000 sq. kilometers of regions within mainland U.S and Bangladesh. We also leveraged citizen science by open-sourcing the dataset and hosting an open competition based on the dataset to rapidly prototype flood extent detection using community generated models. We believe the dataset adds to already existing datasets based on Sentinel-1C SAR data and leads to more robust modeling of flood extents.
  arXiv  Detail & Related papers  (2023-11-15T18:49:29Z)
• DAM-Net: Global Flood Detection from SAR Imagery Using Differential Attention Metric-Based Vision Transformers [22.885444177106873]
  Detection of flooded areas using high-resolution synthetic aperture radar (SAR) imagery is a critical task with applications in crisis and disaster management. To address this issue, we propose a novel differential attention metric-based network (DAM-Net) in this study. The DAM-Net comprises two key components: a weight-sharing Siamese backbone to obtain multi-scale change features of multi-temporal images and tokens containing high-level semantic information of water-body changes.
  arXiv  Detail & Related papers  (2023-06-01T14:12:33Z)
• Towards Daily High-resolution Inundation Observations using Deep Learning and EO [0.0]
  Constantly remote sensing presents a cost-effective solution for synoptic flood monitoring. Satellites do offer timely inundation information when they cover an ongoing flood event, but they are limited by their resolution in terms of their ability to monitor flood evolution at various scales. Data from satellites, such as the Copernicus Sentinels, which have high spatial and low temporal resolution, with data from NASA SMAP and GPM missions could yield high-resolution flood inundation at a daily scale.
  arXiv  Detail & Related papers  (2022-08-10T14:04:50Z)
• Attentive Dual Stream Siamese U-net for Flood Detection on Multi-temporal Sentinel-1 Data [0.0]
  We propose a flood detection network using bi-temporal SAR acquisitions. The proposed segmentation network has an encoder-decoder architecture with two Siamese encoders for pre and post-flood images. The network outperformed the existing state-of-the-art (uni-temporal) flood detection method by 6% IOU.
  arXiv  Detail & Related papers  (2022-04-20T10:56:39Z)
• Flood Segmentation on Sentinel-1 SAR Imagery with Semi-Supervised Learning [1.269104766024433]
  We train an ensemble model of multiple UNet architectures with available high and low confidence labeled data. This assimilated dataset is used for the next round of training ensemble models. Our approach sets a high score on the public leaderboard for the ETCI competition with 0.7654 IoU.
  arXiv  Detail & Related papers  (2021-07-18T05:42:10Z)
• RescueNet: Joint Building Segmentation and Damage Assessment from Satellite Imagery [83.49145695899388]
  RescueNet is a unified model that can simultaneously segment buildings and assess the damage levels to individual buildings and can be trained end-to-end. RescueNet is tested on the large scale and diverse xBD dataset and achieves significantly better building segmentation and damage classification performance than previous methods.
  arXiv  Detail & Related papers  (2020-04-15T19:52:09Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.