GeoSolver: Scaling Test-Time Reasoning in Remote Sensing with Fine-Grained Process Supervision

2026-03-10

AI Summary

This work addresses the susceptibility of existing vision-language models to visual hallucinations and the lack of fine-grained process supervision in multi-step reasoning over remote sensing imagery. To this end, we introduce Geo-PRM-2M, a large-scale dataset with process-level annotations, and propose GeoPRM, a token-level process reward model. We further design Process-Aware Tree-GRPO, a reinforcement learning algorithm that integrates entropy-guided Monte Carlo tree search with a visual hallucination injection mechanism to optimize verifiable reasoning trajectories. Our approach is the first to combine fine-grained process supervision with tree-structured reinforcement learning, enabling test-time reasoning expansion and demonstrating cross-model generalizability. The resulting GeoSolver-9B achieves state-of-the-art performance on multiple remote sensing benchmarks, while GeoPRM significantly enhances both reasoning fidelity and the general capabilities of off-the-shelf vision-language models.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

While Vision-Language Models (VLMs) have significantly advanced remote sensing interpretation, enabling them to perform complex, step-by-step reasoning remains highly challenging. Recent efforts to introduce Chain-of-Thought (CoT) reasoning to this domain have shown promise, yet ensuring the visual faithfulness of these intermediate steps remains a critical bottleneck. To address this, we introduce GeoSolver, a novel framework that transitions remote sensing reasoning toward verifiable, process-supervised reinforcement learning. We first construct Geo-PRM-2M, a large-scale, token-level process supervision dataset synthesized via entropy-guided Monte Carlo Tree Search (MCTS) and targeted visual hallucination injection. Building upon this dataset, we train GeoPRM, a token-level process reward model (PRM) that provides granular faithfulness feedback. To effectively leverage these verification signals, we propose Process-Aware Tree-GRPO, a reinforcement learning algorithm that integrates tree-structured exploration with a faithfulness-weighted reward mechanism to precisely assign credit to intermediate steps. Extensive experiments demonstrate that our resulting model, GeoSolver-9B, achieves state-of-the-art performance across diverse remote sensing benchmarks. Crucially, GeoPRM unlocks robust Test-Time Scaling (TTS). Serving as a universal geospatial verifier, it seamlessly scales the performance of GeoSolver-9B and directly enhances general-purpose VLMs, highlighting its remarkable cross-model generalization.

Problem

• remote sensing
• visual faithfulness
• test-time reasoning
• chain-of-thought
• process supervision

Innovation

• Process Supervision
• Test-Time Scaling
• Vision-Language Models
• Reinforcement Learning
• Remote Sensing