Large Scale Asset Detection Within Railway Scene Point Cloud Data From Mobile Laser Scanning

To reduce greenhouse gas emissions from the transport sector, shifting to rail transport is crucial. This transition will increase the demand on existing rail infrastructure, necessitating large-scale monitoring to maintain its resilience. Point cloud data are an ideal candidate for this purpose, as...

Full description

Saved in:
Bibliographic Details
Main Authors: Bram Ton, Rick Akster
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11086595/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850071957554331648
author Bram Ton
Rick Akster
author_facet Bram Ton
Rick Akster
author_sort Bram Ton
collection DOAJ
description To reduce greenhouse gas emissions from the transport sector, shifting to rail transport is crucial. This transition will increase the demand on existing rail infrastructure, necessitating large-scale monitoring to maintain its resilience. Point cloud data are an ideal candidate for this purpose, as they provide immediate, precise 3D geometric information independent of illumination conditions. This study investigates two object detection models, the PointPillar and the CenterPoint model, to automatically create a digital representation of the rail environment. Using a custom open dataset, these two models are evaluated to detect masts, tension rods, signals, and relay cabinets. A mean Average Precision (mAP@0.5) of 70.6% is achieved. A unique contribution of this study is an in-depth analysis of the locational error in terms of the x and y components of the detected positions. This analysis reveals that location accuracy is not yet sufficient for engineering applications. The analysis indicates that the largest contribution to this error originates from the random error. Additionally, this study demonstrates that transfer learning effectively reduces the labeling burden. For instance, when using 25% of the training data, the average Precision (AP) for the tension rod class improves from 9.5% without transfer learning to 70.8% with transfer learning.
format Article
id doaj-art-e54dfb19dc77489da5007a2eb3b2efd5
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-e54dfb19dc77489da5007a2eb3b2efd52025-08-20T02:47:10ZengIEEEIEEE Access2169-35362025-01-011312911412912610.1109/ACCESS.2025.359077911086595Large Scale Asset Detection Within Railway Scene Point Cloud Data From Mobile Laser ScanningBram Ton0https://orcid.org/0000-0002-9525-5633Rick Akster1Ambient Intelligence Group, Saxion University of Applied Sciences, Enschede, The NetherlandsStrukton Rail, Utrecht, The NetherlandsTo reduce greenhouse gas emissions from the transport sector, shifting to rail transport is crucial. This transition will increase the demand on existing rail infrastructure, necessitating large-scale monitoring to maintain its resilience. Point cloud data are an ideal candidate for this purpose, as they provide immediate, precise 3D geometric information independent of illumination conditions. This study investigates two object detection models, the PointPillar and the CenterPoint model, to automatically create a digital representation of the rail environment. Using a custom open dataset, these two models are evaluated to detect masts, tension rods, signals, and relay cabinets. A mean Average Precision (mAP@0.5) of 70.6% is achieved. A unique contribution of this study is an in-depth analysis of the locational error in terms of the x and y components of the detected positions. This analysis reveals that location accuracy is not yet sufficient for engineering applications. The analysis indicates that the largest contribution to this error originates from the random error. Additionally, this study demonstrates that transfer learning effectively reduces the labeling burden. For instance, when using 25% of the training data, the average Precision (AP) for the tension rod class improves from 9.5% without transfer learning to 70.8% with transfer learning.https://ieeexplore.ieee.org/document/11086595/Deep learningLiDaRmobile laser scanningobject detectionpoint cloudPointPillar
spellingShingle Bram Ton
Rick Akster
Large Scale Asset Detection Within Railway Scene Point Cloud Data From Mobile Laser Scanning
IEEE Access
Deep learning
LiDaR
mobile laser scanning
object detection
point cloud
PointPillar
title Large Scale Asset Detection Within Railway Scene Point Cloud Data From Mobile Laser Scanning
title_full Large Scale Asset Detection Within Railway Scene Point Cloud Data From Mobile Laser Scanning
title_fullStr Large Scale Asset Detection Within Railway Scene Point Cloud Data From Mobile Laser Scanning
title_full_unstemmed Large Scale Asset Detection Within Railway Scene Point Cloud Data From Mobile Laser Scanning
title_short Large Scale Asset Detection Within Railway Scene Point Cloud Data From Mobile Laser Scanning
title_sort large scale asset detection within railway scene point cloud data from mobile laser scanning
topic Deep learning
LiDaR
mobile laser scanning
object detection
point cloud
PointPillar
url https://ieeexplore.ieee.org/document/11086595/
work_keys_str_mv AT bramton largescaleassetdetectionwithinrailwayscenepointclouddatafrommobilelaserscanning
AT rickakster largescaleassetdetectionwithinrailwayscenepointclouddatafrommobilelaserscanning