University of Twente Student Theses
Utilizing Synthetic Point Cloud Generation for Semantic Segmentation of Utility Poles
Dühnen, Julian (2025) Utilizing Synthetic Point Cloud Generation for Semantic Segmentation of Utility Poles.
|
PDF
922kB |
| Abstract: | Machine learning in classification and segmentation of point clouds scanned by LiDAR sensors has been a topic of interest for many years, and has been applied to various fields such as autonomous driving, urban planning and cartography. Despite this, obtaining and labeling real-world ground truth data to train these models remains a time-consuming, error-prone and costly task that is still widely done today, specifically in the field of utility poles and mapping out urban areas. This paper explores the use of synthetic data generation to train a model for semantic segmentation of point clouds of isolated utility poles. This research presents a methodology to generate synthetic point clouds of utility poles and train a model on these synthetic point clouds. The results show that the model trained on synthetic data can be used to segment real-world point clouds with exceptional accuracy, and that procedurally generating synthetic data can be a viable alternative to manually labeling real-world data. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 50 technical science in general, 54 computer science |
| Programme: | Computer Science BSc (56964) |
| Awards: | Best Paper Award |
| Link to this item: | https://purl.utwente.nl/essays/105105 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page