Similarity search for geospatial patches using feature embeddings and vector retrieval

Teacher: Giovanna Venuti

Supervisor: Ing. Gianluca Murdaca – MindEarth

Description: The project aims to develop a prototype for retrieving geospatial areas similar to a given query area based on numerical descriptors or precomputed embeddings. The student will define a suitable spatial unit of analysis, build or use feature representations, and implement a small similarity search workflow capable of returning spatially similar patches according to morphological or land-cover characteristics.

Difficult level: Medium

Requirements: Python, basic GIS, basic machine learning

Scroll to Top