Projective Translations and Affine Stereo Calibration

This paper investigates the homography which transforms a set of points in projective space when undergoing a rigid translation, termed a projective translation. A representation with seven parameters is proposed. They represent explicitly the geometric entities constraining and defining the translation. A practical algebraic method for extracting these parameters is developed. It allows for affine calibration of a stereo rig, for characterizing the motion intrinsics, as well as for composing projective translations. The practical effectiveness of calibration is evaluated on synthetic and real image data.

• Type: Conference
• Year: 1998
• Where: CVPR
• Authors: Andreas Ruf, Gabriela Csurka, Radu Horaud
• Keywords: Computer Vision, CVPR 1998, Practical Algebraic Method, Projective Space, Projective Translations