License
When quoting this document, please refer to the following
URN: urn:nbn:de:0030-drops-18596
URL: http://drops.dagstuhl.de/opus/volltexte/2009/1859/
Go to the corresponding Portal


Munkelt, Christoph ; Denzler, Joachim

Automated Evaluation of 3D Reconstruction Results for Benchmarking View Planning

pdf-format:
Document 1.pdf (135 KB)


Abstract

To obtain complete 3D object reconstructions using optical measurements, several views of the object are necessary. The task of determining good sensor positions to achieve a 3D reconstruction with low error, high completeness and few required views is called the Next Best View (NBV) problem. Solving the NBV problem is an important task for automated 3D reconstruction. However, comparison of different planning methods has been difficult, since only few dedicated test methods exist. We present an extension to our NBV benchmark framework, that allows for faster, automated evaluation of large result data sets. We show that the method introduces insignificant error, while considerably reducing evaluation runtime and increasing robustness.

BibTeX - Entry

@InProceedings{munkelt_et_al:DSP:2009:1859,
  author =	{Christoph Munkelt and Joachim Denzler},
  title =	{Automated Evaluation of 3D Reconstruction Results for Benchmarking View Planning},
  booktitle =	{Computer Vision in Camera Networks for Analyzing Complex Dynamic Natural  Scenes},
  year =	{2009},
  editor =	{Joachim Denzler and Michael Koch},
  publisher =	{Schloss Dagstuhl - Leibniz-Zentrum fuer Informatik, Germany},
  address =	{Dagstuhl, Germany},
  URL =		{http://drops.dagstuhl.de/opus/volltexte/2009/1859},
  annote =	{Keywords: Next Best View, View Planning, Optical 3D Reconstruction, Benchmarking},
}

Keywords: Next Best View, View Planning, Optical 3D Reconstruction, Benchmarking
Seminar: 08422 - Klausurtagung Lehrstuhl Joachim Denzler
Issue Date: 2009
Date of publication: 29.01.2009


DROPS-Home | Fulltext Search | Imprint Published by LZI