EO/OS Object Detection

EO/OS Object Detection (Earth Observation and Overhead Sensing) is a Microsoft first-party computer-vision model that identifies and localizes objects in satellite and aerial imagery, returning bounding-box detections optimized for batch processing of large image archives. Built by the Spectre team behind Microsoft’s Planetary Computer platform, it processes high-resolution remote-sensing data across varied terrain types and atmospheric conditions. The model anchors the new GeoAI category in Microsoft Foundry and is designed to operate at petabyte scale rather than on individual scenes.

The model attacks the chronic bottleneck of manual annotation in environmental monitoring and geospatial intelligence. Automated detection across continental-scale archives enables rapid environmental assessment, infrastructure planning, and change detection in workflows that previously relied on expensive human review. Integration into the Planetary Computer ecosystem extends the capability to researchers and organizations working on climate adaptation, disaster response, and land-use analysis, and is part of Microsoft AI for Science’s broader investment in earth-observation foundation models alongside Aurora.