Operational GeoAI for scalable satellite analytics
Apr 15, 2026 | 3:30 PM - 5:00 PM
Apr 15, 2026 | 3:30 PM - 5:00 PM
Description
Earth Observation (EO) has become a cornerstone of both scientific research and operational services in the space industry. This workshop provides a hands-on, research-oriented introduction to scalable and reproducible GeoAI workflows for satellite imagery, using the MeluXina supercomputer as a concrete execution platform. Participants will implement an end-to-end GeoAI pipeline for large-scale object detection from Sentinel-2 imagery, integrating raster, vector, and auxiliary data sources. The tutorial emphasizes challenges central to current GeoAI research—multimodal learning, spatial generalization, benchmarking, and efficiency.
Level: No prior experience with HPC systems is required. Basic familiarity with Python and deep learning concepts is assumed.
