| KEYWORD |
Evaluating NERF Technology for 3D Reconstruction of Vehicle Accident Scenes
Tesi esterna in azienda
Parole chiave COMPUTER VISION, DEEP LEARNING
Riferimenti ENRICO MAGLI
Gruppi di ricerca ICT4SS - ICT FOR SMART SOCIETIES, Image Processing Lab (IPL)
Tipo tesi AZIENDALE
Descrizione This thesis will be carried out in Fondazione Links, see https://linksfoundation.com/en/
This thesis explores the application of Neural Radiance Fields (NeRF) for the reconstruction of vehicle accident scenes. The primary objective is to assess the capability of NeRF in creating accurate representations of the scenes using data from various sensors installed in vehicles. These sensors may include cameras (e.g. front view, side view) and lidars (e.g. front-facing solid-state LiDAR, 360° rotating LiDAR).
In the envisioned scenario, on-board sensors from surrounding cars are used to capture the scene when they transit in the vicinity of an accident location.
The thesis aims to replicate this scenario on a smaller scale to evaluate the performance and effectiveness of NeRF technology in this specific context.
The study will involve using sensor data collected under controlled conditions simulating a vehicle accident and then using NeRF to reconstruct the scene using the acquired data. The outcomes will provide insights into the feasibility and accuracy of using NeRF for real-world accident scene reconstruction, contributing to the fields of emergency response, insurance, and traffic management.
Conoscenze richieste Basic deep learning skills, including Python and usage of Pytorch or Tensorflow
Scadenza validita proposta 26/01/2025
PROPONI LA TUA CANDIDATURA