< Back to previous page

Project

Efficient 3D reconstruction using mobile devices

In this project, we aim to reconstruct 3D environments using cameras available on mobile devices. This can be used in a wide range of applications, such as 3D project planning, training using AR/VR techniques, movies, gaming, etc. In this project, we will focus on a real estate application. We will start our work from existing 3D reconstruction algorithms based on techniques like structure from motion (SfM), multi-view stereo (MVS) and dense image matching (DIM). We will evaluate these techniques for our application of 3D reconstruction of houses. We will analyze both the quality of the results, computational complexity, and robustness in adverse conditions such as low lighting or large homogeneous surfaces. Next, we will improve on existing state-of-the-art, using techniques such as semantic labeling and object recognition to increase robustness.
Date:24 Oct 2018 →  30 Sep 2020
Keywords:3D reconstructie
Disciplines:Sensors, biosensors and smart sensors, Other electrical and electronic engineering