Are VLMs Lost Between Sky and Space? LinkS$^2$Bench for UAV-Satellite Dynamic Cross-View Spatial Intelligence
This addresses the gap in evaluating VLMs for emergency response and security operations by linking UAV and satellite data, though it is incremental as it builds on existing VLM benchmarks.
The paper tackles the problem of evaluating Vision-Language Models (VLMs) for dynamic cross-view spatial intelligence between UAVs and satellites, introducing LinkS$^2$Bench, a benchmark with 17.9k question-answer pairs, and finds that VLMs significantly lag behind human baselines, with a Cross-View Alignment Adapter improving performance.
Synergistic spatial intelligence between UAVs and satellites is indispensable for emergency response and security operations, as it uniquely integrates macro-scale global coverage with dynamic, real-time local perception. However, the capacity of Vision-Language Models (VLMs) to master this complex interplay remains largely unexplored. This gap persists primarily because existing benchmarks are confined to isolated Unmanned Aerial Vehicle (UAV) videos or static satellite imagery, failing to evaluate the dynamic local-to-global spatial mapping essential for comprehensive cross-view reasoning. To bridge this gap, we introduce LinkS$^2$Bench, the first comprehensive benchmark designed to evaluate VLMs' wide-area, dynamic cross-view spatial intelligence. LinkS$^2$Bench links 1,022 minutes of dynamic UAV footage with high-resolution satellite imagery covering over 200 km$^2$. Through an LMM-assisted pipeline and rigorous human annotation, we constructed 17.9k high-quality question-answer pairs comprising 12 fine-grained tasks across four dimensions: perception, localization, relation, and reasoning. Evaluations of 18 representative VLMs reveal a substantial gap compared to human baselines, identifying accurate cross-view dynamic alignment as the critical bottleneck. To alleviate this, we design a Cross-View Alignment Adapter, demonstrating that explicit alignment significantly improves model performance. Furthermore, fine-tuning experiments underscore the potential of LinkS$^2$Bench in advancing VLM adaptation for complex spatial reasoning.