Title :
Polarization images: a geometric interpretation for shape analysis
Author_Institution :
Dept. of Comput. Sci., Freiburg Univ., Germany
Abstract :
Presents a framework for the geometric interpretation of a single polarization image taken of specular reflecting objects. The task of recovering 3D shape information of specular objects from intensity images is a difficult if not impossible one, as no intensity based features are available. Polarization analysis provides additional features to overcome the problem. In particular the orientation of polarized light is measured, from which constraints on the object surface can be deduced. We show how to perform shape analysis from a single polarization image. The method is applied to local deformation measuring. Along with a discussion of polarization and related geometric issues, algorithms are presented and substantiated by experiments
Keywords :
computer vision; geometry; image reconstruction; light polarisation; object recognition; 3D shape information; geometric interpretation; local deformation; polarization analysis; polarization images; shape analysis; shape recovery; specular reflecting objects; Image analysis; Image reconstruction; Lighting; Optical polarization; Optical reflection; Performance analysis; Rough surfaces; Shape; Surface reconstruction; Surface roughness;
Conference_Titel :
Pattern Recognition, 2000. Proceedings. 15th International Conference on
Conference_Location :
Barcelona
Print_ISBN :
0-7695-0750-6
DOI :
10.1109/ICPR.2000.903602