Existing Q-Learning-based path planning methods face significant bottlenecks in large-scale collaboration, dynamic interference adaptation, and regional value differentiation, failing to meet the practical needs of mountain search and rescue. This study proposes HGR-QL, an optimized Q-Learning method for large-scale multi-UAV operations. Referencing remote sensing datasets, a 50 × 50 dynamic grid environment is constructed by integrating 20% fixed obstacles and 10 moving interference sources, highly simulating real mountain features. Integrating the individual Q-tables and the regional shared Q-tables, the hierarchical independent Q-table architecture is designed, balancing local autonomy and global collaboration. To guide UAVs focusing on remote sensing-identified high-value areas, an innovative multi-level gradient collision avoidance reward function is constructed, avoiding task deviation. Comparative experiments across three scenarios with four baselines and ablation tests validate the core modules. Results show HGR-QL outperforms peers in key metrics: in the dynamic interference scenario, it achieves a 74.47% task completion rate, 25.44 collisions, and a stable 100.00 ms communication delay. HGR-QL provides a lightweight, scalable solution, effectively enhancing the efficiency, safety, and stability of mountain search and rescue and supporting the “golden 72 h” rescue window.

In multi-UAV mountain search and rescue scenarios, the perception system of multi-UAV suffers from low utilization of noise resources, poor collaboration of multi-modal data, and a persistent imbalance between speed and detection accuracy. The paper proposes a federated multi-modal perception method based on terrain-adaptive variational positive-incentive noise (FedVPN). The framework transforms complex mountain interference into task-related beneficial noise, constructs a privacy-preserving federated multi-modal collaborative architecture for distributed feature fusion, and adopts a two-stage training pipeline. Under three typical scenarios, FedVPN outperforms all five baseline methods. In the basic scenario, it achieves an F1-score of 89.23% with a noise gain rate of 7.86%. Under dynamic interference conditions and large-scale heterogeneous environments, the performance decay is only 3.59% and the rescue response time is reduced to 48.60 s. The method significantly improves the accuracy, robustness, and efficiency of the perception module for autonomous rescue systems.

Suitable-area selection reduces candidate-search costs and mismatch risk in road vector map matching. Existing image-based assessments primarily characterize grayscale, texture, and edge saliency, whereas the suitability of road vector maps depends on road continuity, intersection organization, and topological uniqueness. This study proposes a high-confidence suitable-area selection framework based on a Hypergraph-Enhanced Graph Attention Network (HG-GAT). A road-network suitability dataset is constructed using offline matching results. Candidate regions are represented as road graphs, with semantic hyperedges encoding road chains, intersection stars, and key-node neighborhoods. Hypergraph propagation and graph attention jointly learn high-order structural relations and pairwise road connections, while dual-channel readout integrates regional organization and salient local structures. On the Wuhan test data, HG-GAT achieves an accuracy of 0.8074, an F1-score of 0.8309, and a Precision@10% of 0.8980, the highest values for these metrics among the compared methods. It retains high-confidence candidates with a recall of 0.9467 while reducing false selections relative to the image-based and ordinary graph-learning baselines. The results support the value of semantic high-order topology for distinguishing structurally effective road regions and prioritizing candidates for subsequent matching.

Rapid preliminary reconnaissance of Critical Transportation Hubs (CTHs) by UAV swarms is vital during post-disaster rescue operations. However, limited camera fields of view, large heading variances, and GNSS multipath errors near massive steel-concrete structures complicate multi-view cooperative perception. This paper introduces a discrete, vision-based cooperative perception framework utilizing a decentralized anchor-wingman architecture. The pipeline integrates a Prob-IoU-optimized YOLO26m-OBB detector to extract oriented infrastructure footprints. To handle severe rotational discrepancies without IMU priors, a global scene registration cascade—combining SuperPoint and an Optimal Transport-driven LightGlue—is employed to establish robust geometric correspondences. Furthermore, a Projected Polygon Intersection over Union (Proj-IoU) mechanism, coupled with an RMSE-weighted spatial fusion strategy, dynamically associates and deduplicates overlapping targets across distributed views. Experimental results indicate that the framework achieves a low pixel-level RMSE of 2.12 pixels on the source domain and maintains a highly stable 2.36 pixels during zero-shot cross-domain testing (SUES-200 dataset), successfully resolving extreme heading variances up to 270°. The Proj-IoU mechanism resolves multi-source redundancies—collapsing overlapping projections by over 50%—bounding the localization error to approximately 1.06 m. Operating at 6.7 FPS on edge hardware via low-bandwidth tensor transmission, this system provides a rigorous geometric foundation for autonomous swarms, enabling downstream collision-free trajectory planning and Multi-Target Task Allocation (MTTA) in GNSS-denied environments.