Geometric stereo camera calibration with diffractive optical elements

In order to use stereo camera based measurements in machine vision high accuracy geometric camera calibration is absolutely essential. For many applications, e.g. in optical navigation it is necessary to use the epipolar constraint to improve matching algorithms in terms of speed and reliability. Another field of application is the computation of high dense disparity maps where a precise image rectification is needed. The objective is to determine the interior camera parameters including a distortion model as well as the exterior orientation. Therefore, we introduce a two-step approach to calibrate a stereo camera system by means of diffractive optical elements. Working as a beam splitter with precisely defined diffraction angles, it produces a well known diffraction point pattern. As the virtual sources of the diffracted beams are points at infinity, the object to be imaged is invariant against translation. This particular feature allows a complete camera calibration with a single image avoiding complex bundle adjustments, resulting in a very fast and reliable single camera calibration. This procedure has been extended for stereo systems to determine the exterior orientation of both cameras. A comparison with classical photogrammetric methods using chessboard pattern and a calibration with fixed stars exploiting their accurate angular positions will show the capability of the introduced calibration method. The compact calibration setup also allows an in-field calibration.