3DUU
3D Urban Understanding
Through developments in 3D sensor technology, photogrammetry and computer vision, real-world urban scenes can now be captured at a large scale in the form of images or point clouds. This data could support powerful models for applications such as urban planning and self-driving vehicles. However, robustly and efficiently representing these large and dynamic outdoor scenes in a useable semantic 3D representation remains challenging.
In the 3DUU Lab, we will develop new methods and techniques that automatically recognise and model objects in real-world scenes in 3D by combining data from various sources, such as aerial photos and laser scanners on vehicles. We investigate localizing 2D images in the 3D world, reconstructing 3D scenes from such images, and subsequently recognizing objects from 3D or even from multiple sensing modalities simultaneously. Our techniques can thus enrich the data with information about the location and types of objects and surfaces in the scene, such as buildings, streets, trees, traffic lights, and terrains.
The 3DUU Lab is part of the TU Delft AI Labs programme.
Education
Courses
Master Projects
Openings
- Accurate Robot Localization using only Camera Images
Ongoing
- Boriss Bermans: Acoustic Traffic Perception
- Chi Zhang: Multi view diffusion for geometric tasks
- Noortje van der Horst: Procedural Modelling of Tree Growth Using Multi-temporal Point Clouds
- Serenic Monte: Centralized benchmark for 3D vision tasks
- Sharath Chandra Madanu: A deep learning approach to improve the classification of airborne lidar point clouds
- Sitong Li: Enhancing 3D City Models with Neural Representations
- Yingxin Feng: Text-guided geometry editting
Finished
- Alessandro Duico, Alexander Freeman & Boriss Bermans: Can Neural Radiance Fields (NeRF) reconstruct a street from a self-driving vehicle's camera images?
- Camilo Caceres Tocora: Automated Semantic Segmentation of Aerial Imagery using Synthetic Data
- Daniel Dobson: Floor count from street view imagery using learning-based façade parsing
- Fabian Visser: Neural Surface Reconstruction and Stylization
- Ioanna Panagiotidou: Multi View Semantic Reconstruction
- Leon Powalka: Shape-guided artistic route finding
- Linjun Wang: Detailed Facade Reconstruction for Mahattan-world Buildings
- Marios Marinos, Pol Mur i Uribe & Ruben Sangers: Using ensembles of Visual Place Recognition techniques for vehicle localisation
- Wouter de Leeuw: Domain shift-aware Ensemble-based Visual Place Recognition
- Zhaiyu Chen: Learning to Reconstruct Compact Building Models from Point Clouds