MoViS is an internal research project which covers the video exploitation of multi camera systems in vehicles. We aim to prepare the video data in a way that human observers get an intuitive impression of the 3D scene. This can be used either for online navigation tasks or for offline exploitations and comparisons. The more cameras are mounted on a vehicle, the more such an extended representation becomes necessary.
A 3D reconstruction of the scene allows us to generate virtual views in which the observer can navigate.
The project contains the following sub projects:
RobOd MoViS - Robust Odometry Estimation in a Motion based Vision System
MonSter MoViS - Monocular Stereo Vision in a Motion Vision System
MuSiCal MoViS - Multiple and Single Camera Calibration in a Motion based Vision System
DiSCo MoViS - Display and Scene Continuity in a Motion based Vision System