A Hierarchical Geometry-to-Semantic Fusion GNN Framework for Earth Surface Anomalies Detection

The increasing occurrence of earth surface anomalies (ESA) underlines the importance of timely and accurate detection of such events. Therefore, researchers have utilized satellite imagery for large-scale detection and developed advanced deep learning methods. However, the performance is hindered by inadequate labeled data and the complexity of semantic information in satellite imagery. To this end, we propose a hierarchical geometry-to-semantic fusion graph neural network (GNN) framework. Specifically, our method employs two branches to extract geoentities and construct graphs at different levels. Then, a hierarchical graph attention network (GAT) is used to mine complex semantic information from graphs, facilitating accurate and rapid detection of ESA. To fill the gap of the lack of benchmark datasets, we create a composite dataset ESAD based on existing datasets for ESA detection. Extensive experiments demonstrate that the proposed method is effective for accurate ESA detection, outperforming many baseline methods.