GeoScITY proudly announces the successful defense of Abderrazzaq Kharroubi's PhD in July, 2025.
© GeoScITY
Dr. Abderrazzaq Kharroubi’s PhD Defense
Date of Defense: July 15, 2025
Thesis Title: "Semantic segmentation for 3D change detection in urban and railway context using LiDAR point clouds"
Supervisors: Roland Billen (ULiège) and Rafika Hajji (Institut Agronomique et Vétérinaire Hassan II)
Abderrazzaq Kharroubi’s thesis addresses the challenges of 3D change detection in urban and railway environments using multi-temporal LiDAR point clouds. His work develops an object-based framework that fuses semantic segmentation, geometric descriptors, and structured modeling to reliably classify changes such as appeared, disappeared, modified, and unchanged.
The research also introduces a CityJSON-based workflow for updating digital twins and documents infrastructure changes. In addition, Abderrazzaq created the Rail3D LiDAR dataset and applied deep learning to monitor vegetation dynamics in railway networks.
This thesis demonstrates how semantic enrichment and structured outputs can advance automation, reliability, and interpretability in 3D point cloud analysis.
The University of Liège and the GeoScITY Lab congratulate Dr. Kharroubi for his contribution to 3D change detection and digital twin technologies.
Published on 9/2/25
