Automatic road digital twinning from semantically labeled point cloud data

Yuexiong Ding; Mengtian Yin; Ran Wei; Ioannis Brilakis; Muyang Liu; Xiaowei Luo

doi:10.26599/JIC.2026.9180112

Automatic road digital twinning from semantically labeled point cloud data

Yuexiong Ding
, Mengtian Yin
, Ran Wei
, Ioannis Brilakis
, Muyang Liu
, Xiaowei Luo^*

^*Corresponding author for this work

Research output: Journal Publications and Reviews › RGC 21 - Publication in refereed journal › peer-review

5 Downloads (CityUHK Scholars)

Abstract

Creating geometric digital twins (gDTs) for as-built roads still has many limitations, such as low automation level and accuracy, limited asset types and shapes, and reliance on engineering experience. A novel scan-to-building information modeling (scan-to-BIM) framework is proposed for automatic road gDT creation based on semantically labeled point cloud data (PCD), which considers six asset types: road surface, road side (slope), road lane (marking), road/traffic sign, road/street light, and guardrail. The framework first segments the semantic PCD into spatially independent instances or parts, and then extracts the sectional polygon contours as their representative geometric information, stored in JavaScript Object Notation (JSON) files using a new data structure. Primitive gDTs are finally created from the JSON files using the corresponding conversion algorithms. The proposed method achieves an average distance error of 1.46 cm and a processing speed of 6.29 m/s on six real-world road segments with a total length of 1200 m. © The Author(s) 2026.

Original language	English
Article number	9180112
Journal	Journal of Intelligent Construction
Volume	4
Issue number	1
Online published	19 Mar 2026
DOIs	https://doi.org/10.26599/JIC.2026.9180112
Publication status	Published - Mar 2026

Funding

The Shenzhen Science and Technology Innovation Committee (No. JCYJ20180507181647320) and the General Research Fund from the Research Grant Council of Hong Kong (No. SAR 11211622) jointly supported this work. The conclusions herein are those of the authors and do not necessarily reflect the views of sponsoring agencies.

Research Keywords

as-built road
geometric digital twin
geometric information extraction
point cloud data
scan-to-BIM
sectional polygon contour

Publisher's Copyright Statement

This full text is made available under CC-BY 4.0. https://creativecommons.org/licenses/by/4.0/

RGC Funding Information

RGC-funded

Access Output

10.26599/JIC.2026.9180112

460563874.pdfFinal Published version, 4.69 MBLicence: CC BY

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@article{3ef7e7c6d17c49378f8c71afe69781cd,

title = "Automatic road digital twinning from semantically labeled point cloud data",

abstract = "Creating geometric digital twins (gDTs) for as-built roads still has many limitations, such as low automation level and accuracy, limited asset types and shapes, and reliance on engineering experience. A novel scan-to-building information modeling (scan-to-BIM) framework is proposed for automatic road gDT creation based on semantically labeled point cloud data (PCD), which considers six asset types: road surface, road side (slope), road lane (marking), road/traffic sign, road/street light, and guardrail. The framework first segments the semantic PCD into spatially independent instances or parts, and then extracts the sectional polygon contours as their representative geometric information, stored in JavaScript Object Notation (JSON) files using a new data structure. Primitive gDTs are finally created from the JSON files using the corresponding conversion algorithms. The proposed method achieves an average distance error of 1.46 cm and a processing speed of 6.29 m/s on six real-world road segments with a total length of 1200 m. {\textcopyright} The Author(s) 2026.",

keywords = "as-built road, geometric digital twin, geometric information extraction, point cloud data, scan-to-BIM, sectional polygon contour",

author = "Yuexiong Ding and Mengtian Yin and Ran Wei and Ioannis Brilakis and Muyang Liu and Xiaowei Luo",

year = "2026",

month = mar,

doi = "10.26599/JIC.2026.9180112",

language = "English",

volume = "4",

number = "1",

}

TY - JOUR

T1 - Automatic road digital twinning from semantically labeled point cloud data

AU - Ding, Yuexiong

AU - Yin, Mengtian

AU - Wei, Ran

AU - Brilakis, Ioannis

AU - Liu, Muyang

AU - Luo, Xiaowei

PY - 2026/3

Y1 - 2026/3

N2 - Creating geometric digital twins (gDTs) for as-built roads still has many limitations, such as low automation level and accuracy, limited asset types and shapes, and reliance on engineering experience. A novel scan-to-building information modeling (scan-to-BIM) framework is proposed for automatic road gDT creation based on semantically labeled point cloud data (PCD), which considers six asset types: road surface, road side (slope), road lane (marking), road/traffic sign, road/street light, and guardrail. The framework first segments the semantic PCD into spatially independent instances or parts, and then extracts the sectional polygon contours as their representative geometric information, stored in JavaScript Object Notation (JSON) files using a new data structure. Primitive gDTs are finally created from the JSON files using the corresponding conversion algorithms. The proposed method achieves an average distance error of 1.46 cm and a processing speed of 6.29 m/s on six real-world road segments with a total length of 1200 m. © The Author(s) 2026.

AB - Creating geometric digital twins (gDTs) for as-built roads still has many limitations, such as low automation level and accuracy, limited asset types and shapes, and reliance on engineering experience. A novel scan-to-building information modeling (scan-to-BIM) framework is proposed for automatic road gDT creation based on semantically labeled point cloud data (PCD), which considers six asset types: road surface, road side (slope), road lane (marking), road/traffic sign, road/street light, and guardrail. The framework first segments the semantic PCD into spatially independent instances or parts, and then extracts the sectional polygon contours as their representative geometric information, stored in JavaScript Object Notation (JSON) files using a new data structure. Primitive gDTs are finally created from the JSON files using the corresponding conversion algorithms. The proposed method achieves an average distance error of 1.46 cm and a processing speed of 6.29 m/s on six real-world road segments with a total length of 1200 m. © The Author(s) 2026.

KW - as-built road

KW - geometric digital twin

KW - geometric information extraction

KW - point cloud data

KW - scan-to-BIM

KW - sectional polygon contour

UR - https://www.scopus.com/pages/publications/105032786866

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-105032786866&origin=recordpage

U2 - 10.26599/JIC.2026.9180112

DO - 10.26599/JIC.2026.9180112

M3 - RGC 21 - Publication in refereed journal

SN - 2958-3861

VL - 4

JO - Journal of Intelligent Construction

JF - Journal of Intelligent Construction

IS - 1

M1 - 9180112

ER -

Automatic road digital twinning from semantically labeled point cloud data

Abstract

Funding

Research Keywords

Publisher's Copyright Statement

RGC Funding Information

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

GRF: Automatic Detection of Safety Violations using Vision and Knowledge

Cite this

Automatic road digital twinning from semantically labeled point cloud data

Abstract

Funding

Research Keywords

Publisher's Copyright Statement

RGC Funding Information

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Projects

GRF: Automatic Detection of Safety Violations using Vision and Knowledge

Cite this