Adaptive processing of catadioptric images using polarization imaging: towards a pola-catadioptric model

A nonparametric method to define a pixel neighborhood in cata-dioptric images is presented. The method is based on an accurate mod-eling of the mirror shape by mean of polarization imaging. Unlike most processing methods existing in the literature, this method is nonparametric and enables us to respect the catadioptric image's anamorphosis. The neighborhood is directly derived from the two polarization parameters: the angle and the degree of polarization. Regardless of the shape of the catadioptric sensor's mirror (including noncentral configurations), image processing techniques such as image derivation, edge detection, interest point detection, as well as image matching, can be efficiently performed. 1 Introduction Catadioptric sensors are systems consisting of convex mirror and conventional camera [charge coupled device (CCD) or complementary metal-oxide-semiconductor]. 1 The used mirrors can have several different convex shapes, namely, paraboloidal, spherical, hyperbolical, conical, ellipsoidal, or planar. They produce a hemi-spherical field of view.