GeoVLM-R1: Reinforcement Fine-Tuning for Improved Remote Sensing Reasoning

ICLR 2026 Conference Submission17722 Authors

Keywords: Earth Observation, Remote Sensing, GRPO, Reinforcement Learning, VLMs

Abstract: Recent advances in reinforcement learning (RL) have delivered strong reasoning capabilities in natural image domains, yet their potential for Earth Observation (EO) remains largely unexplored. EO tasks introduce unique challenges, spanning referred object detection, image/region captioning, change detection, grounding, and temporal analysis, that demand task-aware reasoning. We propose a novel post-training framework that incorporates task-aware rewards to enable effective adaptation of reasoning-based RL models to diverse EO tasks. This training strategy enhances reasoning capabilities for remote-sensing images, stabilizes optimization, and improves robustness. Extensive experiments across multiple EO benchmarks show consistent performance gains over state-of-the-art generic and specialized vision–language models. Code and models will be released publicly.

Primary Area: reinforcement learning

Submission Number: 17722