2016/04/21 by Andrea Romanoni, Amaël Delaunoy, Romanoni, Andrea +5
Computer Science · Earth and Planetary Sciences · Engineering · #3D Shape Modeling and Analysis #3D Surveying and Cultural Heritage #Advanced Vision and Imaging #Computer Graphics and Visualization Techniques #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences
paper · pdf · doi:10.48550/arxiv.1604.06258
openalex publication_date 2016/04/21 · openalex created_date 2022/10/02 · openalex updated_date 2026/07/28
In this paper we propose a new approach to incrementally initialize a\nmanifold surface for automatic 3D reconstruction from images. More precisely we\nfocus on the automatic initialization of a 3D mesh as close as possible to the\nfinal solution; indeed many approaches require a good initial solution for\nfurther refinement via multi-view stereo techniques. Our novel algorithm\nautomatically estimates an initial manifold mesh for surface evolving\nmulti-view stereo algorithms, where the manifold property needs to be enforced.\nIt bootstraps from 3D points extracted via Structure from Motion, then iterates\nbetween a state-of-the-art manifold reconstruction step and a novel mesh\nsweeping algorithm that looks for new 3D points in the neighborhood of the\nreconstructed manifold to be added in the manifold reconstruction. The\nexperimental results show quantitatively that the mesh sweeping improves the\nresolution and the accuracy of the manifold reconstruction, allowing a better\nconvergence of state-of-the-art surface evolution multi-view stereo algorithms.\n