Semi-Autonomous Camera Calibration of a Stereo Rig for Large Scale Virtual Reality

Practicum

About This Topic

Optical tracking is one of the fundamentals of every virtual reality systems and is based on accurately calibrated cameras. However, camera calibration remains a challenge in case of large camera baselines, large tracking volumes, interferences and insufficient ambient illumination. The main task is to implement a pipeline that calibrates a stereo camera pair based on fusing different visual features. State of the art algorithms to estimate the camera parameters are integrated and the system is tested for accuracy and robustness under different illumination settings.

A stereo camera system, calibration apparatus as well as a number of the pipeline's software modules are available to start the development process. The existing software modules need be extended and integrated into the pipeline. A solid background in programming is necessary, background in multi-view geometry and computer vision is an asset. The topic is ideal to gain a deeper understanding in multi-view geometry, computer vision and optical tracking.

Programming: Matlab, C/C++

Note: this topic can also be started as a practicum and extended to be completed as master thesis.