Uppsats

Automatisk generering av transportytor för 3D-stadsmodeller enligt 3CIM-specifikationen

Master-uppsats

Lunds universitet/Institutionen för naturgeografi och ekosystemvetenskap

Publicerad: 2025

Språk: Svenska

Sammanfattning

This thesis explores the possibilities of integrating semantic transport information in the context of 3D-city modelling by implementing transport surfaces according to standards defined by the 3CIM specification, an extension of CityGML. This approach aims to automatically generate transport surfaces in LOD1 and LOD2, based on the Stockholm city basemap in combination with road geometries that represent roads taken from NVDB. In extension, it also aims to divide the road network into sections and intersections based on CityGML3, which further expands the 3CIM information model. By generating these transport surfaces, they can act as the foundation for digital twins of cities and therefore have plenty of areas of usage. A central goal was to evaluate whether a generalized, repeatable and automatic method could be developed. The process makes use of one training area and two test areas with varying data quality. Results show that the approach is capable enough to produce useful outputs even from low-quality input data. Working around poor data, like an incorrect basemap and poorly defined line objects in the NVDB, it still offers key learnings when investigating an automatic process for the topic. Even though the results show potential, a fully automatic process is still unattainable due to large variations in quality in the basemap. The thesis offers several recommendations for both the basemap and the overall process, and concludes that further work on the topic is not only recommended but likely essential.

Information

Författare
Hellsten, Justin
Lärosäte / institution
Lunds universitet/Institutionen för naturgeografi och ekosystemvetenskap
Publiceringsdatum
2025
Uppsatstyp
Master-uppsats
Språk
Svenska

Utforska vidare

Liknande uppsatser

Uppsatser med liknande ämnen och nyckelord.