Three-Dimensional Object Recognition Using N-Dimensional Chain Codes

Recognition of class and aspect angle of a rigid three-dimensional object in free space is facilitated by using N-dimensional chain codes in feature space. In both radar and image recognition systems features have previously been described that are invariant to object rotation about the observers line of sight to the object center of gravity. However, rotations about the remaining two orthogonal axes produce changes of the features and a full description of the object becomes a toroidal surface in N-dimensional feature space. N-dimensional chain codes are introduced in this paper to give a line approximation of such a surface that is easily used to identify an unknown object and to estimate its aspect angle.