Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors

Images of Building Facades – How Useful Is It for Artifical Intelligence?

Content

The “floor detection” use case aims to address a significant data gap in urban areas: the number of floors in buildings. This information is particularly lacking for older structures, and manually recording it for all buildings in a city would be highly laborious and time-consuming.In this lecture, T4 will present the experimental research process for the automated recording of storeys, from the initial concept to the prototype, and will discuss data acquisition and processing, current challenges, and the results achieved. Another point of discussion will be the question of the extent to which full automation is possible in the use case using AI or whether the results of the AI need to be checked.

Icon

Bilder von Gebäudefassaden - Wie viel steckt für künstliche Intelligenz darin?

Präsentation zur CUT-Akademie am 24.08.2023

The Key Learnings

Acquiring Intelligence Through Artificial Intelligence

The integration of façade images and artificial intelligence facilitates the capture of missing information on buildings with a relatively high degree of certainty.

The More Information, the Greater the Success

Additional information, such as the building type, has been shown to improve the accuracy of predictions regarding the number of storeys.

Adjustments Are Required for the Transfer To Other Cities

The transfer to other cities is possible, but requires a data review and, if necessary, an adaptation of the neural network.

Contact

Aruscha Kramm

Center for Scalable Data Analytics and Artificial Intelligence (ScaDS.AI), Institute of Informatics, Leipzig University

cut@leipzig.de
The Partner Cities
Funded by