This thesis focuses on multi-camera 3D object recognition for intelligent intersections, using multiple cameras mounted on traffic infrastructure to recognise road users. While automated driving is evolving, reliability in complex situations such as large intersections remains a challenge. The VALISENS research project aims to detect the environment more reliably by utilising sensor data from the infrastructure through multi-perspective sensor fusion.
This thesis analyses different approaches to camera-based 3D object recognition for automated driving, in particular using image data from multiple cameras. A comprehensive literature review is conducted to identify existing methods of camera-based 3D object recognition in the context of automated driving. Selected approaches are implemented, trained and evaluated on suitable data sets.
Depending on the scope of the work, it is possible to develop your own approach, which will be included in the comparison. The aim is to evaluate the effectiveness and practicability of different approaches to multi-camera 3D object recognition for roadside perception and to identify new ways to improve the reliability and accuracy of environment perception in the context of automated driving.
This thesis focuses on multi-camera 3D object recognition for intelligent intersections, using multiple cameras mounted on traffic infrastructure to recognise road users. While automated driving is evolving, reliability in complex situations such as large intersections remains a challenge. The VALISENS research project aims to detect the environment more reliably by utilising sensor data from the infrastructure through multi-perspective sensor fusion.
This thesis analyses different approaches to camera-based 3D object recognition for automated driving, in particular using image data from multiple cameras. A comprehensive literature review is conducted to identify existing methods of camera-based 3D object recognition in the context of automated driving. Selected approaches are implemented, trained and evaluated on suitable data sets.
Depending on the scope of the work, it is possible to develop your own approach, which will be included in the comparison. The aim is to evaluate the effectiveness and practicability of different approaches to multi-camera 3D object recognition for roadside perception and to identify new ways to improve the reliability and accuracy of environment perception in the context of automated driving.
Trusting working atmosphere
Exciting and varied projects
Inspiring expert culture
Personal development
Family friendliness
Diversity
XITASO is the expert for high-end software engineering in future-proof digitalisation projects.
We develop individual software solutions of the highest quality for companies facing complex technological and organisational challenges.
We also drive innovation, develop digital strategies as consultants and are a trusted partner on the joint solution path.
We support our customers holistically, with a human touch and at eye level. We want to "enable" them and share our success story with them.
As agile natives, we are convinced that self-organisation, transparency and continuous development are essential prerequisites for entrepreneurial success in our dynamic world.
With enthusiasm and passion, we inspire our environment and lead ourselves and our customers to top performance with excellent results that bring joy.
Enjoy digital excellence.
English 
German 
