Estimation of Surface Normal Vector Based on Majority Voting using Multi-Light Photometric Stereo

Abstract

In this paper, we present an algorithm for estimating surface normal vectors of object from images using multi-light source photometric stereo method in the presence of highlights and shadows. The conventional three-light photometric stereo can not deal with specularity and shadowing because of its simple assumption of a Lambertian surface and local shading. By high intensity or low intensity values, the specular reflection and shadowing will produce an incorrect surface normal in the traditional photometric stereo. We propose a new strategy to eliminate the effects of specularity and shadowing using an algorithm of majority voting in the multi-light source photometry.