A model-based approach to recognition of glossy objects is presented. Normals of surface patches are obtained by analysing the polarizational state of the observed rays under illumination of light sources of circular polarization. The object is assumed to lie alone in its stable pose on the floor. Solid models are examined to find the one that matches the observed normals. First, a candidate model based on the relative angles between known surface normals is found. Then the translation so that the observed positions of surface normals coincide with those of models in the image is found. Some examples are also presented.
展开▼
