GeoMMBench and GeoMMAgent: Toward Expert-Level Multimodal Intelligence in Geoscience and Remote Sensing

TL;DR

This paper introduces GeoMMBench, a comprehensive benchmark for evaluating multimodal models in geoscience and remote sensing, and proposes GeoMMAgent, a multi-agent system that enhances model performance through specialized tools.

Contribution

The paper presents a new benchmark for domain-specific multimodal AI in geoscience and remote sensing, and introduces GeoMMAgent, a multi-agent framework that improves performance over standalone models.

Findings

GeoMMBench enables rigorous evaluation across diverse RS tasks.

GeoMMAgent outperforms standalone LLMs in geospatial tasks.

Tool-augmented agents are crucial for complex geoscience challenges.

Abstract

Recent advances in multimodal large language models (MLLMs) have accelerated progress in domain-oriented AI, yet their development in geoscience and remote sensing (RS) remains constrained by distinctive challenges: wide-ranging disciplinary knowledge, heterogeneous sensor modalities, and a fragmented spectrum of tasks. To bridge these gaps, we introduce GeoMMBench, a comprehensive multimodal question-answering benchmark covering diverse RS disciplines, sensors, and tasks, enabling broader and more rigorous evaluation than prior benchmarks. Using GeoMMBench, we assess 36 open-source and proprietary large language models, uncovering systematic deficiencies in domain knowledge, perceptual grounding, and reasoning--capabilities essential for expert-level geospatial interpretation. Beyond evaluation, we propose GeoMMAgent, a multi-agent framework that strategically integrates retrieval, perception, and reasoning through domain-specific RS models and tools. Extensive experimental results demonstrate that GeoMMAgent significantly outperforms standalone LLMs, underscoring the importance of tool-augmented agents for dynamically tackling complex geoscience and RS challenges.