Large structures

Valid, comprehensive and – most importantly – up-to-date measurement data is essential for planning and developing cities and urban infrastructure. What is the condition of a city’s bridges, buildings, tunnels and roads? How can roads, development works and entire development areas be planned as efficiently and transparently as possible, and how can the construction process be monitored?

Answering these complex questions and working out urban development strategies – also taking into account the resilience of cities against ever more frequent extreme weather events – requires a very extensive data base.

Digital twins of cities

We develop multimodal sensor systems that capture urban infrastructur in great detail and provide georeferenced data. We use cameras and laser scanners that are mounted on mobile platforms such as drones, special surveying vehicles or regularly circulating means of transport. Depending on the requirements, we include additional sensors such as thermal imaging cameras, sound level meters and lighting sensors. The digital measurement data – camera images and 3D point clouds – are processed for use in geographic information systems (GIS). The long-term goal is to create a digital twin of cities as the ideal basis for planning and managing their physical infrastructure.

Cities are spaces that develop dynamically, which is why comprehensive, and – above all – current, measurement data is needed. For this reason, we work on concepts for robust and compact measurement technologies that can be installed on conventional vehicles. In the future, buses or garbage trucks could double as survey vehicles, ensuring the data basis is up-to-date.

Fast data processing and AI-based data evaluation

The sheer size and complexity of cities results in an enormous amount of data being generated. To keep the data volume to the necessary minimum, our measurement systems always contain solutions for real-time processing and data reduction. We use specially trained artificial neural networks to automatically detect and classify desired objects within the data. What’s more, the data our camera images and measurement data provide go far beyond purely geometric object information. RGB color information and values on shadows or reflection surfaces, which, for example, are relevant for noise distribution or 5G networks, can be derived from the data.

 

Press release / 3.6.2025

Sensor box provides real-time geodata for digital urban planning

A mobile sensor box developed at Fraunhofer IPM collects highly accurate 3D geodata. Mounted on buses, taxis, or garbage trucks, the box allows the urban environment to be recorded at frequent intervals without the need for special measurement vehicles. The data serve as a basis for digital urban twins.

 



 

Press release / 16.12.2024

Increasing safety on transportation routes through efficient tree vitality monitoring

 

In the TreeVitaScan project, researchers from Fraunhofer IPM and the University of Freiburg are testing a comprehensive data model that combines multispectral laser scanning with AI-supported data evaluation. This should make it possible to determine the vitality of individual trees automatically, reliably and across a wide area.

 

Press release / 10.6.2024

Quantitative data for thermal building
renovations

An optical mobile mapping system is set to provide comprehensive data for the planning of energy-efficient refurbishments. Its core element is a multispectral LiDAR sensor that measures the geometry of the building and the thermal properties of windows and façades.

 

Excerpt from annual report 2021/22

Focus:
Digital 3D models

Electricity and fiber-optic networks are currently undergoing massive expansion. We have developed a tool that ensures efficient 3D mapping. Objects such as cables, tubes or pipes are automatically recognized and quantified.

 

KI innovation competition

3D-Hydra – AI-based 3D object recognition for flood simulations

In the 3D-Hydra project, an automated process for 3D object recognition in drone data is being developed. The AI-based data interpretation of the point clouds provides a high-resolution 3D model that is designed to make flooding calculations more accurate and efficient in future.