Embedded Files
The project relies on two main sources of data:
Historical photos: visualized on the web-platform and the AR application overlapped to the actual situation. The selected historical photos depict particular events or a different urban structure.
Historical maps: used for the 4D models' realization, also employing machine learning techniques. In particular, maps representing the urban pattern in four different epochs were selected (1851, 1887, 1908, 1936).
The project uses Machine and Deep learning methods to predict building heights in historical maps, colorize and create 3D effects on historical images. We also use re-photographing solutions to geo-reference historical photos and overlay them with actual images. AR methods will facilitate the access to all these results on smartphones or tables while walking in the city.
Historical maps and machine learning methods for the the realization of 4D city models
Re-photographing solutions implemented into a web platform
Page updated
Report abuse