Kalibrierung eines Multikamerasystems für Schlierenversuche im Freifeld

Abstract

The calibration of a large-scale multi-camera setup used to measure blade-tip-vortices of a helicopter in 3D-space is investigated. Therefore a classical calibration method, using a checkerboard as geometrical reference, and a self-calibration method using only 3D points as reference, are analysed. Synthetic image data is created using the pinhole camera model to compare both methods in a virtual multi-camera setup. It is identified that the checkerboard calibration is far more sensitive to image noise than the point based method. This is due to the maximum length of the checkerboard’s edges which is limited to 2m in this application. A multi-camera experiment is performed to demonstrate the concept of self-calibration. To evaluate the obtained accuracy, a 3D-reconstruction of an object with well known geometry is performed. Furthermore the self calibration method provides calculation of the Fundamental matrices. These are needed to perform 3D-reconstructions of blade tip vortices in flight experiments using the epipolar geometry.