Automatic restoration of polygon models

Abstract

We present a fully automatic technique which converts an inconsistent input mesh into an output mesh that is guaranteed to be a clean and consistent mesh representing the closed manifold surface of a solid object. The algorithm removes all typical mesh artifacts such as degenerate triangles, incompatible face orientation, non-manifold vertices and edges, overlapping and penetrating polygons, internal redundant geometry, as well as gaps and holes up to a user-defined maximum size ρ. Moreover, the output mesh always stays within a prescribed tolerance ε to the input mesh. Due to the effective use of a hierarchical octree data structure, the algorithm achieves high voxel resolution (up to 4096 3 on a 2GB PC) and processing times of just a few minutes for moderately complex objects. We demonstrate our technique on various architectural CAD models to show its robustness and reliability.

Visibility-driven Mesh Analysis and Visualization through Graph Cuts

“…Mesh Repair: Our algorithm performs several steps of a mesh repair framework, but we do not currently address hole filling, a major challenge that is implemented in some previous mesh repair systems, e.g. [2,10]. However, the problem of hole filling is not solved by previous work and it remains an inherently difficult topic.…”

“…On the other hand volumetric methods [2,10] can deal with a larger number of inputs, but they are not able to classify the original geometry. As a result a significant increase in triangles is likely for all models that do not have a nice uniform triangulation.…”

“…Volumetric model repair: In recent years a significant attention has been paid to techniques which repair a polygonal model by using an intermediate volumetric representation [16,10,2]. Nooruddin and Turk [16] convert the model into a volumetric grid.…”

“…For the conversion to the polygonal model he suggests to use dual contouring which is able to better preserve sharp features of the model. Bischoff et al [2] proposed a method which is able to robustly handle interpenetrating geometry or dangling triangles. This method also uses octree based voxelization, which is followed by hole filling using morphological operations and an extension of the dual contouring [11].…”

Solid and Physical Modeling

“…Hence, each edge of a valid mesh M is bounding exactly two triangles and each vertex is bounding a single fan of triangles. Sometimes valid polyhedra, and hence valid triangle meshes may be recovered automatically from a set of faces that do not form a watertight surface [Murali97,Bischoff05]. M divides its complement !M into two half-spaces: the inside i(M) and the outside e(M) of M. As explained earlier, i(M) is the set of points from which rays that avoid the edges and vertices of M stab an odd number of triangles of M. Note that M is the boundary of the solid S=M+i(M) and i(M)=iS.…”