Space Partitioning for Distributed Surface Reconstruction

Information

• Publication Type: Student Project
• Workgroup(s)/Project(s):
  • Rendering and Modeling
• Date: 2021
• Date (Start): 2021
• Date (End): March 2022
• Matrikelnummer: 11909464
• First Supervisor:
  • Diana Marin
  • Stefan Ohrhallinger
  • Michael Wimmer
• Keywords: Surface reconstruction, Spatial subdivision

Abstract

The availability of scanners and, implicitly, access to scanned data sets, have increased considerably in the past years. A good example in this direction is the project Wien gibt Raum which provides a 100 TB point cloud for the entire city of Vienna. This technical evolution drives the need for improved surface reconstruction algorithms that can process massive point clouds efficiently.

We have developed a simple algorithm that can be used to reconstruct surfaces from unstructured point clouds. However, in the presence of unoriented surfaces, like in real-life scans, a variant of this algorithm is better adapted to reconstruct the surface. Even if the current implementation is efficient and can handle large data sets in seconds, it would benefit from spatial optimizations.

Additional Files and Images

Weblinks

No further information available.

BibTeX

@studentproject{brunner_lukas_2022_ssb,
  title =      "Space Partitioning for Distributed Surface Reconstruction",
  author =     "Lukas Brunner",
  year =       "2021",
  abstract =   "The availability of scanners and, implicitly, access to
               scanned data sets, have increased considerably in the past
               years. A good example in this direction is the project Wien
               gibt Raum which provides a 100 TB point cloud for the entire
               city of Vienna. This technical evolution drives the need for
               improved surface reconstruction algorithms that can process
               massive point clouds efficiently.  We have developed a
               simple algorithm that can be used to reconstruct surfaces
               from unstructured point clouds. However, in the presence of
               unoriented surfaces, like in real-life scans, a variant of
               this algorithm is better adapted to reconstruct the surface.
               Even if the current implementation is efficient and can
               handle large data sets in seconds, it would benefit from
               spatial optimizations.",
  keywords =   "Surface reconstruction, Spatial subdivision",
  URL =        "https://www.cg.tuwien.ac.at/research/publications/2021/brunner_lukas_2022_ssb/",
}