3D surface reconstruction using polynomial texture mapping

Abstract

A lot of research has been conducted on 3D surface reconstruction using multiple images of fixed view point and varying lighting conditions. In this paper, a new image-based modeling technique based on Polynomial Texture Mapping (PTM) is presented. This technique allows automated reconstruction of highly detailed 3D texture-mapped models. The new technique addresses some of the shortcomings of the previous techniques, mainly in output quality and texture mapping. Moreover, it greatly enhances the perception of existing PTMs. A single PTM is used to generate a uniformly lit diffuse map, a normal map and a height map. The three maps are used to reconstruct a 3D texture-mapped surface. Given the huge archives of PTMs maintained by many institutions and museums, this technique will allow building equally large libraries of 3D models that can improve artifact perception and will aid cultural heritage research. © 2013 Springer-Verlag.