Directly to content
  1. Publishing |
  2. Search |
  3. Browse |
  4. Recent items rss |
  5. Open Access |
  6. Jur. Issues |
  7. DeutschClear Cookie - decide language by browser settings

Robust 3D Surface Reconstruction from Light Fields

Vianello, Alessandro

[thumbnail of Dissertation_Alessandro_Vianello.compressed.pdf]
Preview
PDF, English
Download (30MB) | Terms of use

Citation of documents: Please do not cite the URL that is displayed in your browser location input, instead use the DOI, URN or the persistent URL below, as we can guarantee their long-time accessibility.

Abstract

Light field data captures the intensity, as well as the direction of rays in 3D space, allowing to retrieve not only the 3D geometry information, but also the reflectance properties of the acquired scene. The main focus of this thesis is precise 3D geometry reconstruction from light fields, especially on scenes with specular objects.

A new semi-global approach for 3D reconstruction from linear light fields is proposed. This method combines a modified version of the Progressive Probabilistic Hough Transform with local slope estimates to extract ori- entations, and consequently depth information, in epipolar plane images (EPIs). The resulting reconstructions achieve a higher accuracy than local methods, with a more precise localization of object boundaries, as well as preservation of fine details.

In the second part of the thesis the proposed approach is extended to cir- cular light fields in order to determine the full 360° view of target objects. Additionally, circular light fields allow retrieving depth even from datasets acquired with telecentric lenses, a task which is not possible using a linearly moving camera. Experimental results on synthetic and real datasets demon- strate the quality and the robustness of the proposed algorithm, which pro- vides precise reconstructions even with highly specular objects.

The quality of the final reconstruction opens up many possible application scenarios, such as precise 3D reconstruction for defect detection in industrial optical inspection, object scanning for heritage preservation, as well as depth segmentation for the movie industry.

Document type: Dissertation
Supervisor: Jähne, Prof. Dr. rer. nat. Bernd
Date of thesis defense: 13 November 2017
Date Deposited: 20 Dec 2017 14:16
Date: 2017
Faculties / Institutes: The Faculty of Mathematics and Computer Science > Dean's Office of The Faculty of Mathematics and Computer Science
DDC-classification: 004 Data processing Computer science
About | FAQ | Contact | Imprint |
OA-LogoDINI certificate 2013Logo der Open-Archives-Initiative