Innovation experiment based on circular points and Laguerre theorem in computer vision

Abstract

The image coordinates of circular points calculation is one of the key technologies for camera self-calibration in computer vision. The definitions of planar circular points and isotropic line and Laguerre theorem are given firstly. Then according to the projective invariant of central projection, Laguerre theorem and the nature of the equilateral triangle, the image coordinates of vanishing points in sides of triangles and planar circular points are computed. At last an experiment of camera intrinsic parameters being determined based on the circular points is given. In this experiment, according to obtaining three different images and extract coordinates of every feature points are obtained, the image coordinates of vanishing points and circular points are obtained linearly, and camera intrinsic parameters can be determined through Choleskey decomposition. The result of this experiment shows the approach is effective.