[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85372-en":3,"doc-seo-85372-105":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},85372,7971461741311,"Ophelia","https://ap-avatar.wpscdn.com/avatar/74000253aff267980c6?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779345379180704826",8,"Research & Report","HASTE: A Platform for Rapid Post-Disaster Building Damage Assessment","When a large disaster strikes, responders need an accurate building-level damage map within hours. Traditional machine-learning approaches often rely on matched pre/post imagery and event-relevant training data, which are rarely available on day one. HASTE (High-speed Assessment and Satellite Tracking for Emergencies) is a no-code web platform enabling non-expert analysts to generate per-building damage maps from post-disaster satellite imagery using two lightweight, event-specific workflows. Preliminary xBD experiments show strong separation of damaged vs intact buildings with far fewer labels.","arXiv :2607 . 11838v1 [ cs .CV] 13 Jul 2026  \nMicrosoft (2026), 1–16  \nTECHNICAL REPORT  \nAI for Good  \nLab  \nHASTE: A Platform for Rapid  \nPost-Disaster Building Damage Assessment  \nCaleb Robinson†,* Anthony Ortiz†,* Simone Fobi Nsutezo, Cameron Birge, Meygha Machado, Marcelo Duarte, Joaquin Rivero Rodriguez, Anthony Cintron Roman, Kevin White, Inbal Becker  \nReshef, and Juan M. Lavista Ferres Microsoft AI for Good Research Lab †These authors contributed equally.  \n*Corresponding author. Email: [caleb.robinson@microsoft.com](caleb.robinson@microsoft.com); [anthony.ortiz@microsoft.com](anthony.ortiz@microsoft.com)  \nAbstract  \nWhen a large disaster strikes, responders need a map of which buildings are damaged within hours. The models that do well on public benchmarks assume matched before-and-after imagery and a training set drawn from similar past events, and neither is usually available for a new disaster in its first day. We present HASTE (High-speed Assessment and Satellite Tracking for Emergencies), a no-code web platform that lets analysts who are not machine learning engineers produce per-building damage maps from post-disaster satellite imagery. HASTE implements two methods that share one interface. The first requires the user to label polygons over the post-disaster scene, trains a small semantic segmentation model on that single scene, runs it over the whole image, and joins the per-pixel output to existing building footprints. The second embeds every footprint with a pretrained vision model, requires the user to label a handful of buildings, and fits a logistic regression in the browser that scores the rest ofthe scene in seconds. We describe the platform, both methods, and the engineering that supports them. We also report preliminary experiments on xBD showing that foundation-model embeddings pooled over footprints separate damaged from intact buildings using post-disaster imagery alone, matching a fully supervised ResNet-50 baseline with a twentieth of its labels. HASTE and its predecessors have supported more than thirty real-world disaster responses since 2023, spanning earthquakes, hurricanes, cyclones, floods, wildfires, and tornadoes, delivering results to humanitarian partners within hours to days of imagery becoming available. We close with the directions we think are most promising, including vision-language assessment, active learning, and damage models for roads and other infrastructure. HASTE is open source at [https://github.com/microsoft/haste](https://github.com/microsoft/haste).  \n1. Introduction  \nEarthquakes, hurricanes, floods, and wildfires damage or destroy buildings faster than any ground survey can count them. In the first hours after an event, humanitarian organizations decide where to send search-and-rescue teams, how much shelter to stage, and how to route aid, and these decisions depend on knowing which neighborhoods took the worst damage (Ceferino et al. 2024) . Satellite and aerial imagery can cover a whole city in a single pass, so it has become the main source of this early picture. Turning that imagery into a building-by-building damage map, quickly and reliably, is the problem HASTE addresses.  \nMachine learning has made steady progress on damage assessment, driven largely by the xBD dataset and the xView2 challenge (Gupta, Hosfelt, et al. 2019; DIUx-xView 2019) . The strongest models read a pre-event and a post-event image of the same place and predict a damage level for every pixel or building (Weber and Kané 2020; Adriano et al. 2021; Kaur et al. 2023) . This works well when the test event resembles the training events and when clean imagery from both dates  \n2 Caleb Robinson† et al.  \nFigure 1. HASTE in use. Top, the interactive labeler over the 2023 Lahaina wildfire scene: the analyst has labeled 48 buildings, and the in-browser model scores every footprint in view (green intact, red damaged) alongside holdout metrics for the damaged class. Bottom, the results visualizer, with","cbCaihcAKvWUftz6","https://ap.wps.com/l/cbCaihcAKvWUftz6","pdf",9983249,1,16,"English","en",105,"# Introduction\n## Core Problem and Constraints in Early Disaster Response\n## Machine Learning Limitations Without Matched Pre/Post Data\n## Existing Operational Mapping Approaches\n## How HASTE Addresses the Label and Scale Gap\n# Methods Overview\n## Method 1: Polygon Labeling and On-the-Scene Segmentation\n## Method 2: Footprint Embeddings and In-Browser Logistic Regression\n# Experiments and Impact\n## Preliminary Results on xBD\n## Real-World Deployments Since 2023\n# Future Directions","[{\"question\":\"Why is building damage assessment difficult in the first hours after a new disaster?\",\"answer\":\"Early deployments often lack matched pre- and post-disaster imagery and lack event-specific training data. Changes in appearance, sensor conditions, resolution, and viewing angle further reduce model transfer accuracy.\"},{\"question\":\"What is HASTE and who is it designed for?\",\"answer\":\"HASTE is a no-code web platform that lets analysts who are not machine learning engineers produce per-building damage maps from post-disaster satellite imagery.\"},{\"question\":\"How does HASTE generate per-building damage maps using only limited local labeling?\",\"answer\":\"One workflow uses polygon labels on the post-disaster scene to train a small semantic segmentation model for that single image. The other embeds footprints with a pretrained vision model, uses labels for a handful of buildings, and fits a logistic regression in the browser to score the rest in seconds.\"}]",1784202910,40,{"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},"haste-a-platform-for-rapid-post-disaster-building-damage-assessment","",{"@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/haste-a-platform-for-rapid-post-disaster-building-damage-assessment/85372/",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},"Why is building damage assessment difficult in the first hours after a new disaster?","Question",{"text":75,"@type":76},"Early deployments often lack matched pre- and post-disaster imagery and lack event-specific training data. Changes in appearance, sensor conditions, resolution, and viewing angle further reduce model transfer accuracy.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What is HASTE and who is it designed for?",{"text":80,"@type":76},"HASTE is a no-code web platform that lets analysts who are not machine learning engineers produce per-building damage maps from post-disaster satellite imagery.",{"name":82,"@type":73,"acceptedAnswer":83},"How does HASTE generate per-building damage maps using only limited local labeling?",{"text":84,"@type":76},"One workflow uses polygon labels on the post-disaster scene to train a small semantic segmentation model for that single image. The other embeds footprints with a pretrained vision model, uses labels for a handful of buildings, and fits a logistic regression in the browser to score the rest in seconds.","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},[93,97,101,105,110,115,119,122,127,130,134],{"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":28,"slug":118},7,"Healthcare","healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]