作者: Mark L. Moerdler , John R. Kenderl
DOI:
关键词: Surface reconstruction 、 Orientation (computer vision) 、 Computer science 、 Texel 、 Artificial intelligence 、 Texture (geology) 、 Surface (mathematics) 、 Computer vision
摘要: This paper describes an approach which integrates several conflicting and corroborating shape-from-texture methods in a single system. The system uses new data structure, the augmented texel, combines multiple constraints on orientation compact notation for surface patch. texels initially store weighted that are generated by system's independent components. These texture components, run autonomously may parallel, derive any of currently existing approaches e.g. shape-from-uniform-texel-spacing. For each patch texel then potentially inconsistent data, using Hough transform-like method tesselated gaussian spheres, resulting estimate most likely defines patches part same surface, simplifing reconstruction. This knowledge fusion is illustrated information from two different methods, shape-from-uniform-texel-spacing shape-from-uniform-texel-size. demonstrated camera images artificial natural textures.