The aim of camera calibration is to describe the imaging process of 3D points on the 2D camera chip as precisely as possible using mathematical camera models. Camera calibration distinguishes between an intrinsic (the mapping of 3D points in the camera coordinate system to the 2D camera sensor) and an extrinsic calibration (describes the spatial position and orientation of a camera to a reference coordinate system).
A valid calibration is a basic requirement for further signal processing. Usually the calibration parameters are determined offline. However, it has been shown that vibrations during travel or temperature fluctuations change the orientation of the cameras on the vehicle. This invalidates the extrinsic calibration. Especially when calculating disparity images from a stereo camera system, the extrinsic position must be determined very precisely.
The aim of the research is therefore to develop methods for the online calibration of cameras, especially for stereo systems. The goal is to determine changes in the extrinsic calibration while driving and to constantly adjust the calibration. For this purpose, the extrinsic calibration parameters are currently continuously determined in a recursive estimation process.
