Team: City and Society
The City and Society research group develops innovative AI-based analytical methods to extract objective, reliable and scalable information about urban structure and urban growth, the urban environment and the urban population from satellite and aerial imagery.
Urban heat islands
With around 48,000 heat-related deaths across Germany over the last decade, a precise understanding of urban heat is essential for sustainable urban development in the future, on the path towards climate-resilient cities.
We develop innovative AI-based methods to analyse satellite-based measurements of surface temperature at high temporal and spatial resolution.
Urban greenery
Urban vegetation and green spaces can make a valuable contribution to sustainable, climate-resilient urban development: trees help to cool urban air temperatures through shading and evapotranspiration, and green spaces have a positive impact on the health of urban residents.
We develop AI-based methods to map trees across entire cities using aerial and satellite imagery and to quantify their ecosystem services.
Traffic
Traffic in cities is increasing worldwide, leading to higher noise pollution and air pollution.
We develop AI-based methods for detecting traffic objects using traffic cameras, modelling traffic-related noise across Germany, and investigating urban air pollution using satellite-based measurements worldwide.
Complex Urban Systems (Urban Scaling)
In urban complexity research, cities can be understood as metabolic systems, and certain urban characteristics can be described and understood through scaling laws.
We use urban scaling laws to investigate urban characteristics such as air pollution.
Informal Settlements (Slums)
By 2020, approximately 6 billion people will be living in cities, and this massive migration is giving rise to informal settlements (slums) in many parts of the world, with an estimated population of up to one billion people globally.
We develop AI-based methods to map informal settlements worldwide using satellite imagery and are investigating the development of slums.
Methods
Methods for the analysis of remote sensing data are continuously being refined.
We develop and apply methods of machine learning and deep learning (semantic segmentation, object detection, super-resolution).