Measurement and 3D reconstruction method for pipeline's joints based on shape feature

Tian Zhang, Shao Li Liu*, Cheng Tong Tang, Jian Hua Liu

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

3 Citations (Scopus)

Abstract

To resolve the problem of joint's synchronous measurement and reconstruction in complex pipeline system, a measurement and 3D reconstruction method for pipeline's joints based on shape feature was proposed. A shape feature matching between contours of joint's CAD model and those of joint's material object was adopted to realize the measurement of joint's spatial pose. Projected images of joint's CAD model were acquired through setting up "sphere of viewpoints" of virtual cameras. These projected images were layered in the pyramidal structure to compose a gallery of matching images, and the joint's edges images acquired from material object were matched with its matching images. The joint's spatial pose was calculated iteratively by the least square method, and the joint's 3D model in the pipeline system was reconstructed. A pipeline multi-camera measurement system was proposed, and the experiments on a pipeline's joint pose measurement and 3D reconstruction were carried out. The result showed that the measurement and reconstruction time could be controlled within 1 minute, the position and rotation error were respectively 0.654 mm and 0.73°. The efficiency and accuracy of the proposed measurement and reconstruction method could meet the engineering requirements.

Original languageEnglish
Pages (from-to)123-134
Number of pages12
JournalJisuanji Jicheng Zhizao Xitong/Computer Integrated Manufacturing Systems, CIMS
Volume21
Issue number1
DOIs
Publication statusPublished - 1 Jan 2015

Keywords

  • 3D reconstruction
  • Computer aided design model
  • Pipeline joint
  • Pose measurement
  • Shape feature

Fingerprint

Dive into the research topics of 'Measurement and 3D reconstruction method for pipeline's joints based on shape feature'. Together they form a unique fingerprint.

Cite this