A Study on the High-Precision 3D Digital Modeling Process and Data Optimization Techniques for Port Facilities Based on Terrestrial LiDAR

Maintenance systems for port and fishing port infrastructures are rapidly shifting toward digital transformation and digital twinbased asset management. In this context, terrestrial LiDAR (TLS) has emerged as a key technology for acquiring high-precision three-dimensional digital assets. However, TLS data acquisition in coastal environments is challenging due to tidal variations and the coexistence of fixed and floating structures. This study proposes a practical workflow for high-precision 3D digital modeling and point-cloud data optimization of port facilities using TLS. Field experiments were conducted at Gungpyeong Port using a Leica RTC360 scanner across a 300 m section with 32 scanning stations. To improve reliability under dynamic marine conditions, a threestage registration strategy—field alignment, office precision registration, and dynamic-object segmentation followed by re-registration—was applied. The registration error was reduced from 45.2 mm to 4.3 mm RMSE after the final refinement. In addition, tidal observations confirmed vertical displacement of floating structures of up to 895 mm during the scanning period. For practical use in CAD/BIM environments, a 10 mm uniform sampling technique was applied, reducing the data volume from 6.6 GB to 1.1 GB (83% reduction) while maintaining geometric accuracy within 9.8 mm. The proposed workflow provides a reliable foundation for digital-twin-based monitoring, safety inspection, and intelligent maintenance of port infrastructures.