Extraction of pipes and flanges from point clouds for automated verification of pre-fabricated modules in oil and gas refinery projects

• Published in: Automation in construction Vol. 103; pp. 150 - 167
• Main Authors: Maalek, Reza, Lichti, Derek D., Walker, Ryan, Bhavnani, Adam, Ruwanpura, Janaka Y.
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.07.2019; Elsevier BV
• Subjects: Automated fabrication verification; Automation; Data acquisition; Flanges; Modules; Oil and gas; Pipes; Program verification (computers); Refineries; Robust statistics; Scanners; Terrestrial laser scanner; Robust statistics; Automated fabrication verification; Oil and gas; Terrestrial laser scanner
• ISSN: 0926-5805; 1872-7891
• DOI: 10.1016/j.autcon.2019.03.013

The application of terrestrial laser scanners for fabrication verification of the components of pre-fabricated modules (such as pipes and flanges) is growing markedly in the oil and gas industry. However, there remains strong reliance on impractical and error-prone manual or semi-automated methods to extract semantic information from the acquired point clouds. This manuscript presents a generic and robust framework for automatic extraction of pipe and flange pairs in pre-fabricated modules using the geometric primitives of the point cloud. It has been tested on two point cloud datasets with different data quality and density acquired from different sites. Our method was able to extract all 49 pipes and flanges correctly and improved the accuracy of the estimated centers and normal vectors by 171% and 145%, respectively, when compared to results from commercially-available verification software. The experiments' results show great promise for generic applicability of the proposed system for fabrication verification purposes. [Display omitted] •Development of an automated framework to detect pipes and flanges from point clouds of pre-fabricated oil and gas modules.•Proposed system tested on two point cloud datasets, acquired from separate sites with different data quality and density.•Proposed robust circle extraction method superior compared to popular FAST-LTS in contaminated datasets.•Accuracies of the centre and orientation of flanges improved 171% and 145% compared to commercial verification software.