Lidar Scanning Data Processing Methodology for Creating Detailed Three-Dimensional Models of Transport Infrastructure Objects

The article proposes a methodology for processing digital geospatial data of transport infrastructure objects (on the D5 highway section, Czech Republic) acquired using LiDAR terrain scanning technology (mobile laser scanning system Trimble MX-2) for the creation of three-dimensional digital
models.

The methodological tools for obtaining digital data in the form of point clouds using this system combine the sequential application of specialized methods from geodesy, geoinformatics, and digital cartography (observation, description, geoinformation analysis, and cartographic modeling).

Data processing to achieve the research objectives was carried out based on automatic classification with selective manual editing by a specialist, followed by structuring and optimization, which included the stages of georeferencing, point cloud downsampling to reduce computational load while maintaining sufficient detail, and conversion of the point cloud into a universal format supported by various software applications.
The point cloud processing methodology was built upon fundamental vectorization principles applied during the practical part of the study. These include automatic alignment of objects with the Earth’s surface, prevention of vertical perpendicular lines, polygon closure, prioritization of line representation for unambiguous delineation of vectorized objects, and ensuring vectorization levels corresponding to the elevation characteristics of terrain elements.
The creation of three-dimensional digital models of transport infrastructure objects based on LiDAR scanning for multipurpose practical use represents one approach to developing a spatial data infrastructure using an inductive approach.