Thesis Code: 24006

Thesis type: M.Sc. in Machine Learning, Data Science, Computer Science, Mathematics, or similar

Keywords: 3D data, deep learning, computer vision, aerial imagery


  • Knowledge of Python
  • Software development skills
  • Basic concepts of data science, concerning data analysis, processing and machine learning
  • Basic concepts of image processing, modern computer vision preferred but not required.


The growing interest in achieving enhanced urban planning and management has sparked a surge in exploring advanced technologies for comprehensive data analysis, particularly the utilization of drones for data acquisition in urban environments. This thesis proposal aims to address the segmentation challenges inherent in processing 3D data acquired from drones operating in urban contexts.

The primary focus of this research is to develop advanced segmentation techniques for extracting meaningful information from 3D point clouds obtained by drones. The integration of drone-acquired data into urban planning and environmental management necessitates the development of robust segmentation algorithms that can identify and categorise various urban elements such as buildings, roads, vegetation, and other structures.

The proposed methodology involves several phases. Firstly, a comprehensive review of state-of-the-art techniques in 3D data segmentation will be conducted to identify the most effective methodologies and tools in the field, as well as external datasets of urban point cloud data. Subsequently, the most promising methodologies will be applied to external datasets together with the collected 3D data in Turin.

The evaluation and optimization of the proposed segmentation models will be a crucial aspect of this research. Quantitative metrics, such as accuracy and IoU (interception over union), will be employed to assess the model performance, and adjustments will be eventually made to enhance their effectiveness.

In conclusion, the segmentation of 3D data acquired from drones in urban environments represents a critical step towards unlocking the full potential of drone technology in urban planning and environmental monitoring. This research seeks to contribute novel insights and methodologies to address the challenges associated with extracting meaningful information from drone-acquired 3D data in complex urban landscapes.

Contact: send a resume with attached the list of exams to and specifying the thesis code and title.