H2O-Net: Self-Supervised Flood Segmentation via Adversarial Domain Adaptation and Label Refinement

TL;DR

H2O-Net is a self-supervised deep learning approach that improves flood segmentation accuracy in satellite imagery by bridging domain gaps and refining labels without manual annotations, outperforming existing methods.

Contribution

The paper introduces H2O-Net, a novel self-supervised method for flood segmentation that leverages domain adaptation and label refinement without requiring manual annotations.

Findings

H2O-Net achieves 10-12% higher accuracy than state-of-the-art methods.

The model generalizes well from satellite to drone imagery.

Self-supervision effectively generates high-quality ground truth data.

Abstract

Accurate flood detection in near real time via high resolution, high latency satellite imagery is essential to prevent loss of lives by providing quick and actionable information. Instruments and sensors useful for flood detection are only available in low resolution, low latency satellites with region re-visit periods of up to 16 days, making flood alerting systems that use such satellites unreliable. This work presents H2O-Network, a self supervised deep learning method to segment floods from satellites and aerial imagery by bridging domain gap between low and high latency satellite and coarse-to-fine label refinement. H2O-Net learns to synthesize signals highly correlative with water presence as a domain adaptation step for semantic segmentation in high resolution satellite imagery. Our work also proposes a self-supervision mechanism, which does not require any hand annotation, used during training to generate high quality ground truth data. We demonstrate that H2O-Net outperforms the state-of-the-art semantic segmentation methods on satellite imagery by 10% and 12% pixel accuracy and mIoU respectively for the task of flood segmentation. We emphasize the generalizability of our model by transferring model weights trained on satellite imagery to drone imagery, a highly different sensor and domain.