A SPHERICAL REPRESENTATION FOR RECOGNITION OF FREE-FORM SURFACES

We introduce a new surface representation for recognizing curved objec ts. Our approach begins by representing an object by a discrete mesh o f points built from range data or from a geometric model of the object . The mesh is computed from the data by deforming a standard shaped me sh, for example, an ellipsoid, until it fits the surface of the object . We define local regularity constraints that the mesh must satisfy. W e then define a canonical mapping between the mesh describing the obje ct and a standard spherical mesh. A surface curvature index that is po se-invariant is stored at every node of the mesh. We use this object r epresentation for recognition by comparing the spherical model of a re ference object with the model extracted from a new observed scene. We show how the similarity between reference model and observed data can be evaluated and we show how the pose of the reference object in the o bserved scene can be easily computed using this representation. We pre sent results on real range images which show that this approach to mod elling and recognizing 3D objects has three main advantages: 1) First, it is applicable to complex curved surfaces that cannot be handled by conventional techniques. 2) Second, it reduces the recognition proble m to the computation of similarity between spherical distributions; in particular, the recognition algorithm does not require any combinator ial search. 3) Finally, even though it is based on a spherical mapping , the approach can handle occlusions and partial views.