Landslide disasters are complex spatiotemporal phenomena. Existing deep learning (DL) models for remote sensing (RS) image analysis primarily exploit shallow visual features, inadequately incorporating critical geological, geographical, and environmental knowledge. This limitation impairs detection accuracy and generalization, especially in complex terrains and diverse vegetation conditions. We propose Knowledge Graph Embedding in Swin Feature Pyramid Networks (KGE–SwinFpn), a novel RS landslide segmentation framework that integrates explicit domain knowledge with deep features. First, a comprehensive landslide knowledge graph is constructed, organizing multi-source factors (e.g., lithology, topography, hydrology, rainfall, land cover, etc.) into entities and relations that characterize controlling, inducing, and indicative patterns. A dedicated KGE Block learns embeddings for these entities and discretized factor levels from the landslide knowledge graph, enabling their fusion with multi-scale RS features in SwinFpn. This approach preserves the efficiency of automatic feature learning while embedding prior knowledge guidance, enhancing data–knowledge–model coupling. Experiments demonstrate significant outperformance over classic segmentation networks: on the Yuan-yang dataset, KGE–SwinFpn achieved 96.85% pixel accuracy (PA), 88.46% mean pixel accuracy (MPA), and 82.01% mean intersection over union (MIoU); on the Bijie dataset, it attained 96.28% PA, 90.72% MPA, and 84.47% MIoU. Ablation studies confirm the complementary roles of different knowledge features and the KGE Block’s contribution to robustness in complex terrains. Notably, the KGE Block is architecture-agnostic, suggesting broad applicability for knowledge-guided RS landslide detection and promising enhanced technical support for disaster monitoring and risk assessment.

To address the challenges of complex task matching, limited semantic representation, and low recommendation efficiency in remote sensing data acquisition for natural disasters, this study proposes a semantic path-guided recommendation method based on a knowledge graph framework. A disaster-oriented remote sensing knowledge graph is constructed by integrating entities such as disaster types, remote sensing tasks, observation requirements, sensors, and satellite platforms. High-order meta-paths with semantic closure are designed to model task–resource relationships structurally. A Meta-Path2Vec embedding mechanism is employed to learn vector representations of nodes through path-constrained random walks and Skip-Gram training, capturing implicit semantic correlations between tasks and sensors. Cosine similarity and a Top-K ranking strategy are then applied to perform intelligent task-driven sensor recommendation. Experiments on multiple disaster scenarios—such as floods, landslides, and wildfires—demonstrate the model’s high accuracy and robust stability. An interactive recommendation system is also developed, integrating data querying, model inference, and visual feedback, validating the method’s practicality and effectiveness in real-world applications. This work provides a theoretical foundation and practical solution for intelligent remote sensing data matching in disaster contexts.