SOLVING THE ORIENTATION-DUALITY PROBLEM FOR A CIRCULAR FEATURE IN MOTION

Circular features have been commonly used in numerous computer vision application areas for three-dimensional (3-D) pose estimation, However , the estimation of such a feature's pose from two-dimensional (2-D) i mage coordinates results in an orientation-duality problem. Namely, al though the 3-D position of the circle's center can be uniquely identif ied, the solution process yields two different feasible orientations, of which only one is the true solution. This duality problem is normal ly addressed through the acquisition of a second image in the case of a ''static'' feature, This solution, however, would not be applicable to ''features in motion.'' In this paper, several methods are presente d for the solution of the orientation-duality problem for circular fea tures that are in motion. The first approach is applicable to those fe atures moving on a 3-D line with constant orientation or to those whic h are moving on a plane with general motion. The second approach relie s on the existence of additional object features, such as points or li nes, which are coplanar to the circular feature. In this case, the cir cular feature can undergo an arbitrary 3-D motion. experimental result s verify the validity of the proposed methods.