Estimate of Ball Pose by Monocular Vision Image
基于单视图像的球体姿态估计

A vision inspection technique is provided, which can be used to estimate the position and orientation of the ball and its center axis hole by its projected image. When the focus length of digital camera is given, with the size of circular and spherical feature is also known, their position and/or orientation can be estimated using monocular vision. Thus, the ball pose is acquired from several spherical and circular features shaped by the contour of the ball and its axis. Besides, with the existence of image noise and fitting error of the projection ellipses, the solution base on sole feature is not concurrent with each other. The non-linear least squares algorithm is used to refine above rough solutions and to improve the estimate precise. Experiments with simulated data as well as real image is also presented to validate the algorithms.