RECOGNITION OF PLANAR OBJECTS IN 3-D SPACE FROM SINGLE PERSPECTIVE VIEWS USING CROSS RATIO

The recognition of planar polygons in 3-D space is an important problem in computer vision. This is partially due to the fact that many 3-D objects contain planar polygons, and their recognition is a prerequisite for determining the object's orientations. An approach for recognition of the planar polygon from a single 2-D perspective projection using the basic viewpoint invariant cross-ratio (CR) is presented. The shape CR-descriptor of a spatial planar polygon, i.e., the CR sequence extracted from its single projective image, is proposed in this paper. The minimum mean square error is used as the matching criterion. Trial examples are given. The approach can be directly applied to the recognition and motion analysis of the 3-D objects that contain planar polygons. © 1990 IEEE