Teixeira, AbelCostelha, HugoNeves, CarlosBento, Luis Conde2025-11-282025-11-282024-06-24A. Teixeira, H. Costelha, C. Neves and L. C. Bento, "Point Cloud Alignment for Deposited Material Assessment in Tunnel Environments," 2024 IEEE International Conference on Engineering, Technology, and Innovation (ICE/ITMC), Funchal, Portugal, 2024, pp. 1-7, doi: 10.1109/ICE/ITMC61926.2024.10794331.979-835036243-5979-8-3503-6244-22693-88552334-315Xhttp://hdl.handle.net/10400.8/14775keywords provenientes de ieee.Conference name 30th ICE IEEE/ITMC Conference on Engineering, Technology, and Innovation, ICE/ITMC 2024, FunchalConference date 24 June 2024 - 28 June 2024The assessment of deposited material in tunnel reinforcement operations can be performed using a 3D model generated from multiple scans. For this purpose, an accurate alignment of the scanned models is required. Aligning existing structure model with data scanned after surface deformations can be challenging, particularly if reference markers are not available or were displaced. For scenarios where the surrounding structure is largely changed, certain procedures can be adapted when processing the scanned data to achieve consistent alignment between scanned and reference structure models. This work proposes a methodology to cope with these situations, analysing the impact of different approaches. Experiments were performed in a realistic scenario related with shotcrete of railway tunnels wall surfaces, with the results showing the applicability of the developed work. The proposed procedure relies in highlighting the importance of specific points that describe the same feature in the reference and aligning PC. The proposed methodology achieved an RMS difference of 0.0173 m, which lead to a drastic improvement in the point cloud alignment compared to the use of standard ICP algorithm without data preprocessing, which achieved 0.0518 m in the studied use-case.engPoint cloud compressionAdaptation modelsTechnological innovationThree-dimensional displaysRail transportationData modelsTrajectorySurface treatmentRobotsStandardsPoint cloudRegistrationTunnel scanningRobotic shotcreteconference paper2025-11-27cv-prod-443771810.1109/ice/itmc61926.2024.10794331