SimStad is a multipart project with an overall goal of collecting and visualizing data of cities for city maintainance and planning, and public outreach.
Classification using deep learning
Municipals spend a lot of time and resources each year collecting data from their cities. By using deep learning semantic segmentation (AI), some of this data can automatically be converted into GIS-data (Geographic Information System), leaving more time and resources for other work.
The quality of drone and aircraft photos is not always enough. By using deep learning, images from drones and aircrafts can be improved by for example increasing the resolution, removing certain objects, or removing shadows.
Exhibition at Norrköping Visualization Center C
A large 3D-print of Norrköping has been created as a part of an exhibit at Visualization Center C. Projectors can show various datasets on the 3D-model, such as flooding risks and noise levels in the city.
A digital twin of Norrköping will be created as a tool to better understand the city.