3D reconstruction method of bent tube with NURBS curve

Tian Zhang*, Jianhua Liu, Chengtong Tang, Shaoli Liu

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

8 Citations (Scopus)

Abstract

To solve the difficulty and low efficiency of three-dimensional reconstruction of complex spatial form bent tube, a stereo vision method with NURBS curve to achieve three-dimensional reconstruction of bent tube is presented. This method first extracts central axis and the edge line of bent tube by the grayscale image, and the outer diameter of bent tube is calculated through the edge line extracted. On the premise of affine camera model, NURBS curve has been chosen as the primitive of reconstruction, and by using its affine invariance, according to the principle of binocular vision, the central axis of bent tube is reconstructed through reconstructing the control points. Finally, the three-dimensional model of bent tube is reconstructed by three-dimensional modeling technique. Experimental results show that, relative to the traditional methods, this method makes a big improvement on the efficiency and accuracy of three-dimensional reconstruction of bent tube.

Original languageEnglish
Title of host publicationMechatronics and Applied Mechanics II
Pages112-118
Number of pages7
DOIs
Publication statusPublished - 2013
Event2nd International Conference on Mechatronics and Applied Mechanics, ICMAM 2012 - , Taiwan, Province of China
Duration: 8 Dec 20129 Dec 2012

Publication series

NameApplied Mechanics and Materials
Volume300-301
ISSN (Print)1660-9336
ISSN (Electronic)1662-7482

Conference

Conference2nd International Conference on Mechatronics and Applied Mechanics, ICMAM 2012
Country/TerritoryTaiwan, Province of China
Period8/12/129/12/12

Keywords

  • Bent tube
  • NURBS curve
  • Stereo vision
  • Three-dimensional reconstruction

Fingerprint

Dive into the research topics of '3D reconstruction method of bent tube with NURBS curve'. Together they form a unique fingerprint.

Cite this