Automated georegistration of 3D models
Master semester project at Digital Humanities Laboratory (EPFL)
Author: Alicia Soria Gómez
Supervisor: Albane Descombes
Year: Autumn semester 2021
Introduction This project explores different approaches to achieve automatic geolocalization of points in 3D models.
- Image matching
- 3D to 2D matching
- Synthesized view matching
Dependencies The main libraries employed are: ee, folium, cv2, google_streetview, geopy, streetview_functions, defisheye, numpy, plotly.express, matplotlib
Results The repo includes the Python Notebooks that served as training
- finding features in OpenCV.ipynb
- geolocalization_addresses_geopy.ipynb
- rolexandvortex_project.qgz
- try_to_download_panoramics.ipynb
And those with the final results of each approach:
- OpenMVG training.docx
- aerial_matching.ipynb
- camera_movement_1.ipynb
- gross_localization_to_panoramic.ipynb
- musee_louvre.qgz