Incremental Solid Modeling from Sparse Structure-from-Motion Data with Improved Visual Artifacts Removal

Abstract

Abstract-In the recent years, a family of 2-manifold surface reconstruction methods from a sparse Structure-from-Motion points cloud based on 3D Delaunay triangulation was developed. This family consists of batch and incremental variations which include a step that remove visual artifacts. Although been necessary in the term of surface quality, this step is slow compared to the other parts of the algorithm and is not well suited to be used in an incremental manner. In this paper, we present two other methods for removing visual artifacts. They are evaluated and compared to the previous one in the incremental context where the need of new methods is the highest. Taken separately, they provide medium results, but used together they are as good as the old method in the terms of surface quality, and at the same time, processing time is almost three times smaller.

Real-Time CPU-Based Large-Scale Three-Dimensional Mesh Reconstruction

“…For this reason Litvinov and Lhuiller [1] proposed an algorithm to update incrementally the mesh while keeping the manifold property valid in all the iterations. The growing procedure, in that algorithm, carves the 3D space sometimes resulting in visual artifacts that [29] where the same authors proposed an ad-hoc method to explicitly remove visual artifacts and in [2] where the authors changed the ordering of the carving procedure in order to preemptively remove problematic tetrahedra, i.e., those most likely to result in visual artifacts.…”

Mesh-based 3D textured urban mapping

“…Voxel-based approaches usually produce unappealing reconstructions, due to the voxelization of the space, and they need a very high resolution to capture fine details of the scene, trading off their efficiency. In Computer Vision community, different volumetric representations have been explored, in particular many algorithms adopt the 3D Delaunay triangulation [13], [5], [6], [14]. Delaunay triangulation is self-adaptive according to the density of the data, i.e., the points, without any indexing policy; moreover its structure is made up of tetraedra from which it is easy to extract a triangular mesh, widely used in the Computer Graphics community to accurately model objects.…”

Overview of Shelling for 2-Manifold Surface Reconstruction Based on 3D Delaunay Triangulation

“…The first one (on the top) is inspired by the real example in Fig. 1 of [16] such that M is "the matter" (set of the tetrahedra non-intersected by rays and including the ground and walls of a city), P , S and P ∪ S are 3-balls that cannot be carved by a shelling although they are in the free-space (set of the tetrahedra intersected by rays). The second one (on the bottom) is similar and is inspired by the real example in Fig.…”

“…It also introduces examples of visual artifact similar to that in Fig. 1 of [16] which cannot be removed by shelling alone (against the intuition).…”

“…There are both batch [13] and incremental [23,15,16,18] variations. First a 3D Delaunay triangulation is computed from the sparse point cloud to divide up the space by a set T of tetrahedra.…”