Automatic surface patch generation from a video image sequence

Yixiang Tian, M. Gerke, G. Vosselman, Qing Zhu

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

4 Citations (Scopus)
1 Downloads (Pure)


Automatic 3D building reconstruction is becoming increasingly important for a number of applications. This paper aims to present a new approach for an automatic pipeline to reconstruct surface patches from a video image sequence, which can deal with images acquired with an uncalibrated hand held camera. In order to skip the details to get the surface patches that present the building’s main structure, the algorithm applies rules and a processing sequence that aim to reasonably group extracted sparse 3D points and 3D edges. Then the surface patches are estimated from these grouped points and edges. The final step is to define the surface patches’ outline, which aim to both present the location and the actual shape of the surface patches. The results for two datasets show that the main surface patches for each dataset are successfully reconstructed.
Original languageEnglish
Title of host publication3D Geo-Information Sciences
EditorsJiyeong Lee, Sisi Zlatanova
Place of PublicationBerlin
Number of pages12
ISBN (Print)978-3-540-87394-5
Publication statusPublished - 2009
Event3rd International Workshop 3D Geo-Information 2008 - Seoul, Korea, Republic of
Duration: 13 Nov 200814 Nov 2008
Conference number: 3

Publication series

NameLecture Notes in Geoinformation and Cartography
ISSN (Print)1863-2246
ISSN (Electronic)1863-2351


Workshop3rd International Workshop 3D Geo-Information 2008
Country/TerritoryKorea, Republic of


  • ADLIB-ART-330
  • EOS
  • 2023 OA procedure


Dive into the research topics of 'Automatic surface patch generation from a video image sequence'. Together they form a unique fingerprint.

Cite this