Automatic analysis of geographic areas of interest using Vision-Language Models on multi-layer geospatial data

Teacher: Giovanna Venuti

Supervisor: Ing. Gianluca Murdaca – MindEarth

Description: The project aims to develop a prototype for the automatic description of geographic Areas of Interest (AOIs) based on multi-layer geospatial inputs. The student will work with visual representations combining satellite imagery and derived layers, such as built-up maps, change maps, or simple indicators, in order to generate structured textual descriptions supporting the interpretation of the territorial context.

Difficult level: Medium – High

Requirements: Python, basic GIS and geospatial data handling,
interest in multimodal AI, basic deep learning

Scroll to Top