MODEL-BASED POINT CLOUD RECONSTRUCTION of EXCAVATION SITE for BURIED PIPE MANAGEMENT

Kouki Kurita, Masafumi Nakagawa

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Construction space modeling is required for automated operation of construction vehicles, construction management, and safety management. Construction sites are constantly changing during drilling work. Thus, a base map of the construction site should be updated using point clouds acquired through several steps in construction work. Particularly, multiple buried objects such as underground pipes should be measured in construction work because buried objects are difficult to measure with 3D scanners after such work is complete. Moreover, it is better to visualize buried objects with software products such as virtual reality and augmented reality to avoid accidental destruction because of excavation work using construction vehicles such as backhoes. Therefore, 3D measurement and modeling of buried objects are significant in construction work. In this study, we proposed a rapid 3D measurement and modeling of underground objects in construction work. First, we estimate the central axis of point clouds for the 3D modeling of buried pipes. Second, the 3D pipe model was fit to acquired point clouds. Buried pipes can be represented as known geometry models because of products provided based on industrial standards. Thus, the 3D model fitting approach was applied using a pipe model prepared in advance. Then, after 3D model fitting, the depths of buried pipes from ground surfaces were calculated using the estimated axis and point clouds of ground surfaces. We conducted an experiment on 3D measurement based on point cloud acquisition with SfM/MVS and 3D laser scanner during electric line pipe installation work. We also experimented on the 3D modeling of pipes using point clouds acquired with a handheld camera and 3D laser scanner to reconstruct 3D shapes and positions of pipes at a construction site.

Original languageEnglish
Title of host publication42nd Asian Conference on Remote Sensing, ACRS 2021
PublisherAsian Association on Remote Sensing
ISBN (Electronic)9781713843818
Publication statusPublished - 2021
Event42nd Asian Conference on Remote Sensing, ACRS 2021 - Can Tho, Viet Nam
Duration: 2021 Nov 222021 Nov 26

Publication series

Name42nd Asian Conference on Remote Sensing, ACRS 2021

Conference

Conference42nd Asian Conference on Remote Sensing, ACRS 2021
Country/TerritoryViet Nam
CityCan Tho
Period21/11/2221/11/26

Keywords

  • 3D Modeling
  • BIM
  • Model Fitting
  • Point clouds
  • SfM-MVS

ASJC Scopus subject areas

  • Information Systems

Fingerprint

Dive into the research topics of 'MODEL-BASED POINT CLOUD RECONSTRUCTION of EXCAVATION SITE for BURIED PIPE MANAGEMENT'. Together they form a unique fingerprint.

Cite this