[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85519-en":3,"doc-seo-85519-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},85519,34359740700684,"Finn","https://ap-avatar.wpscdn.com/avatar/1f400023980c374ae676?_k=1777273430885731487",8,"Research & Report","OpenEarthAgent: A Unified Framework for Tool-Augmented Geospatial Agents","OpenEarthAgent introduces a unified framework for tool-augmented geospatial reasoning, extending multimodal agent capabilities from general vision to remote sensing. The approach trains on satellite imagery, natural-language queries, and structured reasoning traces, addressing the need to reason over spatial scale, geographic structures, and multispectral indices while preserving coherent multi-step logic. A unified executable tool registry and trajectory-based policy learning standardize visual, spectral, GIS, and georeferenced raster operations with deterministic replay validation for spatial correctness and executability.","OpenEarthAgent: A Unified Framework for Tool-Augmented Geospatial Agents  \nAkashah Shabbir 1⋆, Muhammad Umer Sheikh 1 ⋆ , Muhammad Akhtar Munir 1 , Hiyam Debary2 , Mustansar Fiaz2 , Muhammad Zaigham Zaheer 1 , Paolo Fraccaro2 , Fahad Shahbaz Khan 1 ,3 , Muhammad Haris Khan 1 , Xiao Xiang Zhu4 , and  \nSalman Khan 1 ,5  \narXiv :2602 . 17665v4 [ cs .CV] 12 Jul 2026  \n1 Mohamed bin Zayed University of AI, United Arab Emirates  \n2 IBM Research, Europe  \n3 Linköping University, Sweden  \n4 Technical University Munich, Germany  \n5 Australian National University, Australia  \n􀂀https://github.com/mbzuai-oryx/OpenEarthAgent  \nFig. 1: Comparison of OpenEarthAgent-4B and Qwen3-4B on a complex GIS reasoning task. OpenEarthAgent correctly sequences tool calls with proper dependencies and feedback, while Qwen3 fails due to misordered tool usage and inconsistent reasoning.  \nAbstract. Recent progress in multimodal reasoning has enabled agents that interpret imagery, connect it with language, and execute structured analytical tasks. Extending these capabilities to remote sensing remains challenging, as models must reason over spatial scale, geographic structures, and multispectral indices while maintaining coherent multistep logic. To address this gap, we introduce OpenEarthAgent, a unified framework for tool-augmented geospatial reasoning trained on satellite imagery, natural-language queries, and structured reasoning traces. Beyond serving as a benchmark, OpenEarthAgent establishes a cohesive agentic architecture built around a unified executable tool registry and trajectory-based policy learning. The framework standardizes heterogeneous visual, spectral, GIS, and georeferenced raster operations under  \n⋆ Equal contribution.  \n2 A. Shabbir et al.  \na consistent callable schema, enabling modular orchestration and deterministic execution. Training is performed via supervised fine-tuning on structured reasoning trajectories with deterministic replay validation to ensure executability and spatial correctness. The accompanying corpus comprises 14,538 training and 1,169 evaluation instances with over 107K reasoning steps, spanning urban, environmental, disaster, and infrastructure domains and incorporating GIS 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 tool-driven behavior across diverse EO scenarios. We report consistent improvements over a strong baseline and competitive performance against recent open and closed-source models.  \n1 Introduction  \nThe evolution of visual representation learning has transitioned from static perception to interactive multimodal reasoning. Early single-shot frameworks such as DINO [4] and MAE [14] laid the foundation for large-scale visual understanding through self-supervised objectives, learning strong image encoders without explicit language-based reasoning or interaction. Inspired by these successes, the remote sensing community extended this paradigm to earth observation, where models such as Prithvi [17], Copernicus-FM [43], Galileo [39], Panopticon [40], TerraFM [7], and AnySat [1] scaled representation learning to global multisensor data. These earth-scale vision models demonstrated remarkable transfer across spatial resolutions and modalities, yet their predictions remained one-shot and perception-centric, focusing on recognition rather than structured reasoning.  \nIn parallel, large multi-modal models have progressed beyond perception. Architectures such as BLIP-2 [20], InstructBLIP [6], LLaVA-OneVision [19], and Kosmos-2 [30] extended language models into the visual domain, enabling grounded image understanding. Building on this trajectory, remote-sensing VLMs emerged to adapt multimodal reasoning to geospatial data. Early efforts such as RemoteCLIP [23], SkySenseGPT [25], and GeoChat [18] introduced largescale multimodal alignment for remote sensi","cbCaiqbOAMnhPkcF","https://ap.wps.com/l/cbCaiqbOAMnhPkcF","pdf",4954473,1,39,"English","en",105,"# Abstract\n# Introduction","[{\"question\":\"What problem does OpenEarthAgent target in remote sensing agents?\",\"answer\":\"OpenEarthAgent targets the gap in structured, tool-driven reasoning for geospatial tasks, where current remote-sensing models often remain descriptive and lack explicit reasoning with spatially verifiable outputs.\"},{\"question\":\"How is OpenEarthAgent trained and validated for executability?\",\"answer\":\"Training uses supervised fine-tuning on structured reasoning trajectories, and deterministic replay validation ensures the tool calls are executable and spatially correct.\"},{\"question\":\"What data and tool capabilities does the framework support?\",\"answer\":\"The dataset includes training and evaluation instances with reasoning traces spanning urban, environmental, disaster, and infrastructure domains, integrating GIS operations and index analyses such as NDVI, NBR, and NDBI through a unified callable tool schema.\"}]",1784204146,98,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"openearthagent-a-unified-framework-for-tool-augmented-geospatial-agents","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/openearthagent-a-unified-framework-for-tool-augmented-geospatial-agents/85519/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does OpenEarthAgent target in remote sensing agents?","Question",{"text":75,"@type":76},"OpenEarthAgent targets the gap in structured, tool-driven reasoning for geospatial tasks, where current remote-sensing models often remain descriptive and lack explicit reasoning with spatially verifiable outputs.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How is OpenEarthAgent trained and validated for executability?",{"text":80,"@type":76},"Training uses supervised fine-tuning on structured reasoning trajectories, and deterministic replay validation ensures the tool calls are executable and spatially correct.",{"name":82,"@type":73,"acceptedAnswer":83},"What data and tool capabilities does the framework support?",{"text":84,"@type":76},"The dataset includes training and evaluation instances with reasoning traces spanning urban, environmental, disaster, and infrastructure domains, integrating GIS operations and index analyses such as NDVI, NBR, and NDBI through a unified callable tool schema.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]