Accurate Positioning Using Sphere in Single Views

An accurate and flexible positioning method using a sphere is presented, which can be used in machine vision systems. By exploring a new geometric interpretation that the image of a sphere is concentric with the image of the absolute conic, it is shown that there are three constraint equations that can be used for fitting the image of the sphere with the known camera intrinsic parameters. Experimental results demonstrate that the positioning accuracy of this method is much higher than that of the traditional method