# Towards Faithful Reasoning in Remote Sensing: A Perceptually-Grounded GeoSpatial Chain-of-Thought for Vision-Language Models

arXiv: https://arxiv.org/abs/2509.22221 · HTML: https://mlsanity.com/paper/2509.22221/

## Key facts

- Cited by 19 papers in the map (data through 14 Oct 2026).
- Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning.
- Builds on 15 of its 63 references that are in the map.
- First posted on arXiv 26 Sep 2025.

## Placed in

- LLM reasoning / [Spatial and embodied reasoning](https://mlsanity.com/lr/spatial-and-embodied-reasoning/)
- Multimodal LLMs / vision-language models / [Embodied and 3D vision-language models](https://mlsanity.com/mllm/embodied-and-3d-vision-language-models/)

## Builds on (most cited first, 15 of 15)

- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://mlsanity.com/paper/2507.06261/) (2025)
- [MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models](https://mlsanity.com/paper/2304.10592/) (2023)
- [MiniGPT-v2: large language model as a unified interface for vision-language multi-task learning](https://mlsanity.com/paper/2310.09478/) (2023)
- [GLM-4.5V and GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning](https://mlsanity.com/paper/2507.01006/) (2025)
- [Kimi-VL Technical Report](https://mlsanity.com/paper/2504.07491/) (2025)
- [SkySenseGPT: A Fine-Grained Instruction Tuning Dataset and Model for Remote Sensing Vision-Language Understanding](https://mlsanity.com/paper/2406.10100/) (2024)
- [DriveCoT: Integrating Chain-of-Thought Reasoning with End-to-End Driving](https://mlsanity.com/paper/2403.16996/) (2024)
- [SegEarth-R1: Geospatial Pixel Reasoning via Large Language Model](https://mlsanity.com/paper/2504.09644/) (2025)
- [RemoteReasoner: Towards Unifying Geospatial Reasoning Workflow](https://mlsanity.com/paper/2507.19280/) (2025)
- [CoMT: Chain-of-Medical-Thought Reduces Hallucination in Medical Report Generation](https://mlsanity.com/paper/2406.11451/) (2024)
- [CMMCoT: Enhancing Complex Multi-Image Comprehension via Multi-Modal Chain-of-Thought and Memory Augmentation](https://mlsanity.com/paper/2503.05255/) (2025)
- [RingMo-Agent: A Unified Remote Sensing Foundation Model for Multi-Platform and Multi-Modal Reasoning](https://mlsanity.com/paper/2507.20776/) (2025)
- [GeoChain: Multimodal Chain-of-Thought for Geographic Reasoning](https://mlsanity.com/paper/2506.00785/) (2025)

## Built on by (most cited first, 19 of 19)

- [GeoZero: Incentivizing Reasoning from Scratch on Geospatial Scenes](https://mlsanity.com/paper/2511.22645/) (2025)
- [RSGround-R1: Rethinking Remote Sensing Visual Grounding through Spatial Reasoning](https://mlsanity.com/paper/2601.21634/) (2026)
- [RemoteAgent: Bridging Vague Human Intents and Earth Observation with RL-based Agentic MLLMs](https://mlsanity.com/paper/2604.07765/) (2026)
- [Earth-OneVision: Extending Remote Sensing Multimodal Large Language Models to More Sensor Modalities and Tasks](https://mlsanity.com/paper/2606.10819/) (2026)
- [GeoSolver: Scaling Test-Time Reasoning in Remote Sensing with Fine-Grained Process Supervision](https://mlsanity.com/paper/2603.09551/) (2026)
- [RemoteZero: Geospatial Reasoning with Zero Labels](https://mlsanity.com/paper/2605.04451/) (2026)
- [WeaveEarth: Structured Evidence Construction and Reasoning for Training-Free UHR Remote Sensing Understanding](https://mlsanity.com/paper/2607.10120/) (2026)
- [Uni-RS: A Spatially Faithful Unified Understanding and Generation Model for Remote Sensing](https://mlsanity.com/paper/2601.17673/) (2026)
- [GeoX: Mastering Geospatial Reasoning Through Self-Play and Verifiable Rewards](https://mlsanity.com/paper/2605.20006/) (2026)
- [GeoSearcher: Anchor-Guided Progressive Reasoning for Remote Sensing Visual Grounding with Process Supervision](https://mlsanity.com/paper/2607.01050/) (2026)
- [Think and Answer ME: Benchmarking and Exploring Multi-Entity Reasoning Grounding in Remote Sensing](https://mlsanity.com/paper/2603.12788/) (2026)
- [Earth Science Foundation Models: From Perception to Reasoning and Discovery](https://mlsanity.com/paper/2605.12542/) (2026)
- [Beyond Appearance: Can Multimodal Large Language Models Exploit Vertical Structure for Remote Sensing Natural Scene Understanding?](https://mlsanity.com/paper/2605.25784/) (2026)
- [More with Less: a Large Scale Remote Sensing VLM with a Simple Recipe](https://mlsanity.com/paper/2607.15942/) (2026)
- [FUSAR-R1: A Large-Scale Reasoning Model for Intelligent Interpretation of SAR Images](https://mlsanity.com/paper/2607.16819/) (2026)
- [The Earth in One Gaze: Training-Free Active Focus for UHR Remote Sensing Understanding](https://mlsanity.com/paper/2609.31747/) (2026)
- [Aperture: Training-Free Multiscale Concept Bottlenecks for Remote Sensing](https://mlsanity.com/paper/2609.38603/) (2026)
- [Hallucination-aware Faithful Reasoning in Remote Sensing](https://mlsanity.com/paper/2610.18369/) (2026)
- [ReDiS: Reasoning-Diagnosis Self-Evolution for Multimodal Reasoning in Remote Sensing](https://mlsanity.com/paper/2610.18388/) (2026)

## Abstract (arXiv)

Vision-Language Models (VLMs) in remote sensing often fail at complex analytical tasks, a limitation stemming from their end-to-end training paradigm that bypasses crucial reasoning steps and leads to unverifiable outputs. To address this limitation, we introduce the Perceptually-Grounded Geospatial Chain-of-Thought (Geo-CoT), a framework that models remote sensing analysis as a verifiable, multi-step process. We instill this analytical process through a two-stage alignment strategy, leveraging Geo-CoT380k, the first large-scale dataset of structured Geo-CoT rationales. This strategy first employs supervised fine-tuning (SFT) to instill the foundational cognitive architecture, then leverages Group Reward Policy Optimization (GRPO) to refine the model's reasoning policy towards factual correctness. The resulting model, RSThinker, outputs both a final answer and its justifying, verifiable analytical trace. This capability yields dominant performance, significantly outperforming state-of-the-art models across a comprehensive range of tasks. The public release of our Geo-CoT380k dataset and RSThinker model upon publication serves as a concrete pathway from opaque perception towards structured, verifiable reasoning for Earth Observation.