Satellite Land-Use and Urban Growth Mapping

A land-use mapping and urban-growth detection system for Sentinel-2 satellite imagery built around Prithvi-EO-1.0, the geospatial foundation model of IBM and NASA, and trained on a student laptop CPU. On 2,000 held-out EuroSAT-MS patches the ensemble reaches 96.2 % accuracy. Real Sentinel-2 scenes of Madrid, Hyderabad and Bengaluru from 2017, 2021 and 2024 are mapped, urban growth is measured in square kilometres and checked against ESA WorldCover, all delivered as the LandUse Watch web application.

Category: Deep Learning / Remote Sensing / Geospatial AI / Urban Analytics

CSE Project 26 from ProTechSC Academy, with graphical abstract, video demonstration and project guidance.