Geo-R1: Unlocking VLM Geospatial Reasoning with Cross-View Reinforcement Learning
This addresses the challenge of geospatial reasoning for AI applications in domains like mapping and environmental analysis, representing an incremental advancement in model training methods.
The paper tackles the problem of enabling geospatial reasoning in vision-language models by introducing Geo-R1, a post-training framework that uses supervised fine-tuning and reinforcement learning, achieving state-of-the-art performance on various benchmarks.
We introduce Geo-R1, a reasoning-centric post-training framework that unlocks geospatial reasoning in vision-language models by combining thinking scaffolding and elevating. In the scaffolding stage, Geo-R1 instills a ``geospatial thinking paradigm" via supervised fine-tuning on synthetic chain-of-thought exemplars, enabling models to connect visual cues with geographic priors without costly human reasoning annotations. In the elevating stage, it uses GRPO-based reinforcement learning on a weakly-supervised cross-view pairing proxy. This design supplies a verifiable and scalable reward signal: teaching models to capture and reconcile features across modalities, and harnessing reasoning for accurate prediction. Geo-R1 extends geospatial modeling from domain pretraining / supervised finetuning to reasoning-first post-training, and achieves state-of-the-art performance across various geospatial reasoning benchmarks. Our model is available at https://huggingface.co/miniHui/Geo-R1.