Deep Learning for Image-based Localization and Mapping
Title: Deep Learning for Image-based Localization and Mapping
DNr: Berzelius-2024-335
Project Type: LiU Berzelius
Principal Investigator: Viktor Larsson <viktor.larsson@math.lth.se>
Affiliation: Lunds universitet
Duration: 2024-10-01 – 2025-04-01
Classification: 10207
Homepage: https://vlarsson.github.io
Keywords:

Abstract

Many applications require knowing the position and orientation of a device. For example, path-planning in robotics or virtual/augmented reality. For solving this task, cameras are a common senor choice as they can provide accurate estimates while being relatively cheap compared to LiDARs or other active sensors. In this project we will train deep neural networks which will be integrated into image-based localization and mapping pipelines.