Mosaicing for automated pipe scanning

Rahul Summan, Neha Mathur, Gordon Dobie, Graeme West, Stephen Marshall, Carmelo Mineo, Charles Norman MacLeod, Stephen Pierce, William Kerr

Research output: Contribution to conferencePaperpeer-review

95 Downloads (Pure)

Abstract

The interior visual inspection of pipelines in the nuclear industry is a safety critical activity conducted during outages to ensure the continued safe and reliable operation of plant. Typically, the video output by a manually deployed probe is viewed by an operator looking to identify and localise surface defects such as cracks, corrosion and pitting. However, it is very challenging to estimate the nature and extent of defects by viewing a large structure through a relatively small field of view. This work describes a new visual inspection system (Figure 1) employing photogrammetry using a fisheye camera and a structured light system to map the internal geometry of pipelines by generating a photorealistic, geometrically accurate surface model (Figure 2). The error of the system output was evaluated through comparison to a ground truth laser scan of a nuclear grade split pipe sample containing artificial defects representative of the application – the error was found to be submillimetre across the sample.
Original languageEnglish
Number of pages20
Publication statusPublished - 24 Nov 2016
Event1st 3D Metrology Conference - Aachen, Germany
Duration: 22 Nov 201624 Nov 2016

Conference

Conference1st 3D Metrology Conference
Abbreviated title3DMC 2016
Country/TerritoryGermany
CityAachen
Period22/11/1624/11/16

Keywords

  • remote visual inspection
  • 3D reconstruction
  • pipe scanning

Fingerprint

Dive into the research topics of 'Mosaicing for automated pipe scanning'. Together they form a unique fingerprint.

Cite this