Automatic pipe and elbow recognition from three-dimensional point cloud model of industrial plant piping system using convolutional neural network-based primitive classification
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Kim Y. | - |
dc.contributor.author | Nguyen C.H.P. | - |
dc.contributor.author | Choi Y. | - |
dc.date.available | 2020-08-03T05:21:12Z | - |
dc.date.issued | 2020-08 | - |
dc.identifier.issn | 0926-5805 | - |
dc.identifier.issn | 1872-7891 | - |
dc.identifier.uri | https://scholarworks.bwise.kr/cau/handle/2019.sw.cau/42675 | - |
dc.description.abstract | With the recent development of laser scanning technology, the variety of applications of laser scanners has increased. One typical application is object recognition from laser-scanned point cloud models. On large-scale construction sites such as refineries and industrial plants, object recognition from point cloud models has been widely employed for construction progress monitoring, assembly inspections, and maintenance purposes. Pipelines are among the main objects of interest with regard to object recognition on such sites. There has been extensive research on recognizing pipes in pipelines; however, research on recognizing pipe-connecting elbows is still lacking. Most representative elbow recognition methods are centerline-based and connectivity-based methods. These methods do not use laser-scanned points directly; instead, they employ feature values that are calculated from laser-scanned points. However, these feature values are easily affected by noise and occlusion; therefore, the elbow recognition results could be inaccurate owing to noisy and occluded point cloud models. In this paper, we propose an automatic pipe and elbow recognition method robust against noise and occlusion in which pipes and elbows are recognized directly from laser-scanned points. This method starts with pipeline extraction, followed by elbow classification based on curvature information. Falsely classified points are filtered using convolutional neural network-based primitive classification. After elbow recognition is completed, pipe classification and recognition are performed. Experimental results obtained from three different point cloud models demonstrated that the proposed method recognizes pipes and elbows with high accuracy from noisy and occluded point cloud models. © 2020 Elsevier B.V. | - |
dc.language | 영어 | - |
dc.language.iso | ENG | - |
dc.publisher | Elsevier B.V. | - |
dc.title | Automatic pipe and elbow recognition from three-dimensional point cloud model of industrial plant piping system using convolutional neural network-based primitive classification | - |
dc.type | Article | - |
dc.identifier.doi | 10.1016/j.autcon.2020.103236 | - |
dc.identifier.bibliographicCitation | Automation in Construction, v.116 | - |
dc.description.isOpenAccess | N | - |
dc.identifier.wosid | 000539120100024 | - |
dc.identifier.scopusid | 2-s2.0-85084492043 | - |
dc.citation.title | Automation in Construction | - |
dc.citation.volume | 116 | - |
dc.type.docType | Article | - |
dc.publisher.location | 네델란드 | - |
dc.subject.keywordAuthor | Automatic pipe and elbow recognition | - |
dc.subject.keywordAuthor | Convolutional neural network-based primitive classification | - |
dc.subject.keywordAuthor | Curvature | - |
dc.subject.keywordAuthor | Elbow classification | - |
dc.subject.keywordPlus | Classification (of information) | - |
dc.subject.keywordPlus | Cloud computing | - |
dc.subject.keywordPlus | Convolution | - |
dc.subject.keywordPlus | Industrial plants | - |
dc.subject.keywordPlus | Laser applications | - |
dc.subject.keywordPlus | Object recognition | - |
dc.subject.keywordPlus | Pipelines | - |
dc.subject.keywordPlus | Classification and recognition | - |
dc.subject.keywordPlus | Construction progress | - |
dc.subject.keywordPlus | Construction sites | - |
dc.subject.keywordPlus | Curvature information | - |
dc.subject.keywordPlus | Laser scanning technology | - |
dc.subject.keywordPlus | Recognition methods | - |
dc.subject.keywordPlus | Three-dimensional point clouds | - |
dc.subject.keywordPlus | Typical application | - |
dc.subject.keywordPlus | Convolutional neural networks | - |
dc.relation.journalResearchArea | Construction & Building Technology | - |
dc.relation.journalResearchArea | Engineering | - |
dc.relation.journalWebOfScienceCategory | Construction & Building Technology | - |
dc.relation.journalWebOfScienceCategory | Engineering, Civil | - |
dc.description.journalRegisteredClass | scie | - |
dc.description.journalRegisteredClass | scopus | - |
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.
84, Heukseok-ro, Dongjak-gu, Seoul, Republic of Korea (06974)02-820-6194
COPYRIGHT 2019 Chung-Ang University All Rights Reserved.
Certain data included herein are derived from the © Web of Science of Clarivate Analytics. All rights reserved.
You may not copy or re-distribute this material in whole or in part without the prior written consent of Clarivate Analytics.