Chair and Keynote Speaker
Prof. Dr. Stephan Nebiker
FHNW University of Applied Sciences and Arts
Northwestern, Switzerland

Biography
Stephan Nebiker is Professor for Geoinformatics, Photogrammetry and Remote Sensing at FHNW University of Applied Sciences Northwestern Switzerland. His research interests include various aspects of 3D geoinformation technologies such as 3d vision mobile mapping, virtual globes, UAVs, cloud computing, geospatial augemented reality, as well a automated information extraction. Stephan Nebiker is head of the Masters programme in geoinformation technology at FHNW and is co-founder of the FHNW spin-off iNovitas AG.

Abstract
3D Vision Mobile Mapping for Urban and Traffic Infrastructure Management in the Cloud

Recent developments in kinematic positioning, imaging sensors, 3d vision algorithms, and cloud technologies enable radically new, entirely image-based approaches for managing urban or traffic infrastructure. In our presentation we discuss recent developments and the state-of-technology and application in 3d vision mobile mapping and in the web- and cloud-based exploitation of the resulting 3d image databases. First experiments in mobile mapping were based on imaging sensors and date back to the early 80ies. However, over the last decade, the rapidly evolving mobile mapping market has largely been dominated by LiDAR technology. With the establishment of highly popular geo web services, such as Google Street View, in the mass market, the potential of image-based solutions has become apparent to a broad audience, including potential professional users in fields such as infrastructure management. In our research and in the resulting infra3Dâ„¢ technology we employ stereovision mobile mapping and combine it with cloud and web technologies in order to enable high-quality 3d infrastructure management at the desks of the individual domain experts with an unparalleled ease-of-use. First, we introduce the mobile acquisition platform consisting of a GNSS/INS positioning system and up to 5 stereo camera systems. These can be operated at up to 30 fps resulting in dense stereo image sequences even at high driving speeds. We then address new calibration and integrated georeferencing approaches which enable absolute 3D measurement accuracies at the 1 cm level and relative accuracies better than 1 cm. Subsequently, we introduce the cloud-based infra3Dâ„¢ architecture for processing and exploiting massive volumes of geospatial 3D imagery. Among the special features are a flexible integration of the infra3D-cloud services into existing GIS by means of a powerful WebSDK. This permits, for example, the seamless and accurate integration of 2D geodata into the 3D geoinformation services. The remaining part of our presentation is dedicated to use cases from highway, railway and urban infrastructure management in order to discuss the capabilities, strengths and limitations of 3D image-based solutions.