Deep Learning Framework and Performance Evaluation for Point-Cloud Time-Series Change Detection in Long-Term Monitoring of Ancient Bridges
Main Article Content
Abstract
Ancient bridges exhibit progressive deformation and localized spalling due to long-term material weathering, humidity–temperature cycles, foundation settlement, and traffic loading, making accurate structural health monitoring essential for heritage preservation. In advanced electromagnetic and intelligent sensing systems, high-precision geometric monitoring complements radar- and wave-based inspection technologies for comprehensive infrastructure assessment. Conventional geometric-differencing methods for point-cloud change detection are susceptible to accumulated registration errors, noise occlusion, and non-uniform point density during long-term monitoring, limiting reliable identification of millimeter-level deformations. This study establishes a deep learning framework for multi-temporal point-cloud change modeling to satisfy the requirements of life-cycle structural monitoring. The temporal characteristics of structural changes and the propagation mechanisms of registration errors are analyzed, and mathematical formulations for cloud-to-cloud (C2C) and cloud-to-plane (C2P) change metrics are presented. The proposed framework provides a theoretical foundation for deep learning-based change detection and performance evaluation, while offering technical support for multimodal structural sensing and intelligent monitoring applications in engineering infrastructures.
Downloads
Article Details
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
References
H. Zhang, M. Yuan, Q. Jiang, et al., “Deep learning model compression for real-time vibration monitoring of wood-boring activities,” Journal of Beijing Forestry University, vol. 43, no. 6, pp. 9, 2021, doi: 10.12171/j.1000-1522.20200100.
Z. Li, C. Quan, H. Chen, et al., “Building change detection method based on point clouds generated by dense image matching,” Geospatial Information, vol. (3), pp. 022, 2024, doi: 10.3969/j.issn.1672-4623.2024.03.003.
C. Meng, Y. Zhao, W. Han, et al., “Urban building change detection using airborne LiDAR point clouds based on RandLA-Net,” Remote Sensing for Natural Resources, pp. 34(4), 2022, doi: 10.6046/zrzyyg.2021402.
J. Chai, Y. Zhang, Z. Yang, et al., “Multi-Level building change detection using DSM and DOM generated from UAV imagery,” Remote Sensing for Land & Resources, vol. (2), pp. 9, 2024, doi: 10.6046/zrzyyg.2023001.
I. De Gélis, S. Lefèvre, and T. Corpetti, “3D urban change detection with point cloud Siamese networks,” Copernicus GmbH, 2021, doi: 10.5194/ISPRS-ARCHIVES-XLIII-B3-2021-879-2021.
H. D. Seo and E. M. Kim, “Object Classification and Change Detection in Point Clouds Using Deep Learning,” LX Spatial Information Research Institute, 2020, doi: 10.22640/LXSIRI.2020.50.2.37.
G. Kaur and Y. Afaq, “Developments in deep learning for change detection in remote sensing: A review,” Transactions in GIS, pp. 28(2), 2024, doi: 10.1111/tgis.13133.
Léo Renaut, H. Frei, and NüchteAr, “Deep learning on 3D point clouds for fast pose estimation during satellite rendezvous,” Acta Astronautica, vol. 232, pp. 231-243, 2025, doi: 10.1016/j.actaastro.2025.03.009.
J. Li, P. Fearnhead, P. Fryzlewicz, et al., “Automatic change-point detection in time series via deep learning,” Journal of the Royal Statistical Society: Series B (Statistical Methodology), pp. 86(2), 2024, doi: 10.1093/jrsssb/qkae004.
T. D. Ryck, M. D. Vos, and A. Bertrand, “Change Point Detection in Time Series Data Using Autoencoders with a Time-Invariant Representation,” IEEE Transactions on Signal Processing, 2021, doi: 10.1109/TSP.2021.3087031.
H. He, J. Yan, D. Liang, et al., “Time-Series land cover change detection using deep learning-based temporal semantic segmentation,” Remote Sensing of Environment, vol. 305, Art. no. 114101, 2024, doi: 10.1016/j.rse.2024.114101.
T. Han and G. A. Sánchez-Azofeifa, “A Deep Learning Time Series Approach for Leaf and Wood Classification from Terrestrial LiDAR Point Clouds,” Remote Sensing, vol. 14, no. 13, pp. 3157, 2022, doi: 10.3390/rs14133157.
S. Shirowzhan, M. Samad, et al., “Comparative analysis of machine learning and point-based algorithms for detecting 3D changes in buildings over time using bi-temporal lidar data - ScienceDirect,” Automation in Construction, vol. 105, Art. no. 102841, 2019, doi: 10.1016/j.autcon.2019.102841.
A. Floris, L. Frittoli, D. Carrera, et al., “Composite Convolution: A Flexible Operator for Deep Learning on 3D Point Clouds,” Pattern Recognition, Art. no. 110557, 2024, doi: 10.1016/j.patcog.2024.110557.
T. Chaton, N. Chaulet, S. Horache, et al., “Torch-Points3D: A Modular Multi-Task Frameworkfor Reproducible Deep Learning on 3D Point Clouds,” 2020 International Conference on 3D Vision (3DV), 2020, doi: 10.1109/3DV50981.2020.00029.