AI summaryⓘ
The authors studied how language models can analyze very detailed satellite images where important clues are spread out and hard to find. They discovered that just zooming in on small areas helps with easier tasks but not with harder ones that need looking at multiple places or comparing regions. To fix this, they created a big dataset called GeoMTVR with many questions and tools that let models break down complex problems and choose different ways to inspect the images. They also made a new learning method to help models decide the best tools to use and when. Their combined approach, GeoLens, does better than using just zooming or direct reasoning when interpreting large satellite images.
Ultra-high-resolution imageryMultimodal large language models (MLLMs)Zoom-in toolVisual question answering (VQA)Geospatial reasoningReinforcement learningSupervised fine-tuningSatellite imageryTool-use trajectoryEvidence grounding
Authors
Fengxiang Wang, Jiangnan Huang, Mingshuo Chen, Yueying Li, Yang Shi, Junwei Luo, Haoyu Wang, Yansheng Li, Jing Zhang, Haiyan Zhao, Wenjing Yang
Abstract
Ultra-high-resolution (UHR) remote-sensing (RS) imagery provides fine-grained Earth-observation evidence over city-scale scenes, but poses a fundamental challenge for multimodal large language models (MLLMs): task-relevant evidence is often sparse, local, and spatially dispersed across extremely large visual contexts. A natural solution is to equip MLLMs with zoom-in tools for active local inspection. However, through a pilot study on XLRS-Bench, we find that zoom-in is only partially effective: it resolves easy and medium-level tasks with locally recoverable evidence, but saturates on hard cases requiring global search, multi-region comparison, path planning, or dispersed-evidence reasoning. Motivated by this finding, we move beyond single-tool zoom-in and introduce GeoMTVR, a large-scale Geospatial Multi-Tool Visual Reasoning dataset built from wide-area satellite imagery. GeoMTVR contains 13K UHR VQA samples with interleaved reasoning trajectories, diverse visual tool calls, and returned visual observations, enabling models to learn question decomposition, tool selection, regional inspection, object-level grounding, auxiliary visual reasoning, and cross-tool evidence integration. Beyond supervised fine-tuning, we propose a tool-attention-focused reinforcement learning algorithm that concentrates optimization on critical tool-use decisions, including when to invoke tools, which tool to select, where to apply it, and how to interpret tool outputs. By combining SFT on GeoMTVR with our RL algorithm, we develop GeoLens, a multi-tool visual reasoning MLLM for UHR RS. Experiments show that GeoLens consistently outperforms direct reasoning and single-tool zoom-in baselines, achieving stronger accuracy, better evidence grounding, and more efficient tool-use trajectories.