OpenEarthAgent: A Unified Framework for Tool-Augmented Geospatial Agents
2026-02-19 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors developed OpenEarthAgent, a system that helps computers understand and analyze satellite images by combining language and image information in a step-by-step manner. They trained the model using many examples that include detailed reasoning steps and tools used for geographic analysis. Their dataset covers different areas like cities, environment, disasters, and infrastructure, and uses common satellite image indices like NDVI. The trained agent can reason clearly about locations and tasks, showing better performance than strong previous methods. This work aims to make geospatial analysis more interpretable and reliable.
multimodal reasoningremote sensingsatellite imageryGIS (Geographic Information Systems)NDVI (Normalized Difference Vegetation Index)NBR (Normalized Burn Ratio)NDBI (Normalized Difference Built-up Index)tool-augmented agentsstructured reasoningsupervised fine-tuning
Authors
Akashah Shabbir, Muhammad Umer Sheikh, Muhammad Akhtar Munir, Hiyam Debary, Mustansar Fiaz, Muhammad Zaigham Zaheer, Paolo Fraccaro, Fahad Shahbaz Khan, Muhammad Haris Khan, Xiao Xiang Zhu, Salman Khan
Abstract
Recent progress in multimodal reasoning has enabled agents that can interpret imagery, connect it with language, and perform structured analytical tasks. Extending such capabilities to the remote sensing domain remains challenging, as models must reason over spatial scale, geographic structures, and multispectral indices while maintaining coherent multi-step logic. To bridge this gap, OpenEarthAgent introduces a unified framework for developing tool-augmented geospatial agents trained on satellite imagery, natural-language queries, and detailed reasoning traces. The training pipeline relies on supervised fine-tuning over structured reasoning trajectories, aligning the model with verified multistep tool interactions across diverse analytical contexts. The accompanying corpus comprises 14,538 training and 1,169 evaluation instances, with more than 100K reasoning steps in the training split and over 7K reasoning steps in the evaluation split. It spans urban, environmental, disaster, and infrastructure domains, and incorporates GIS-based operations alongside index analyses such as NDVI, NBR, and NDBI. Grounded in explicit reasoning traces, the learned agent demonstrates structured reasoning, stable spatial understanding, and interpretable behaviour through tool-driven geospatial interactions across diverse conditions. We report consistent improvements over a strong baseline and competitive performance relative to recent open and closed-source models.