An Intelligent and Robust System for Underwater Vision

  • Cameras are one of the most common sensors in robotics. They are used in both research and industrial applications. Underwater images suffer from extremely unfavourable conditions. Light is heavily attenuated and scattered. Furthermore, underwater images are distorted due to refraction through water-glass-air interfaces. All of this makes using cameras underwater exceptionally difficult. Work presented in this thesis is mainly motivated by the needs of three EU-founded projects: MORPH, CADDY and DexROV. Work in these projects resulted in development at every stage in the design, calibration and image processing of underwater vision systems The first part introduces investigated research problems, and describes the background and motivation of the thesis. Furthermore, exemplar projects are described to give a good intuition of the state of the art and what advances have been made in this thesis. The second part of this thesis describes the issue of refraction-based distortions. An extensive analysis of the problem is undertaken using a novel method, dubbed Pinax, that allows for very efficient and accurate modelling of submerged cameras. The Pinax model does not require any underwater calibration: a single in-air procedure is sufficient to handle a variety of underwater environments, including different salinities, temperatures and pressures. The third part focuses on designing the stereo vision system from the perspective of selecting hardware and setup parameters that would perform best in a given task. Part IV addresses the problem of image degradation. The image formation process is discussed and an adaptation of the Dark Channel Prior to underwater conditions is proposed. This resulted with an image correction algorithm that allows for a reduction in backscattering in the registered images. Part V describes of the practical applications of the methods presented earlier. The achieved results are discussed in terms of their real-life applications.

Download full text

Cite this publication

  • Export Bibtex
  • Export RIS

Citable URL (?):

Search for this publication

Search Google Scholar Search Catalog of German National Library Search OCLC WorldCat Search Catalog of GBV Common Library Network Search Catalog of Jacobs University Library Search Bielefeld Academic Search Engine
Meta data
Publishing Institution:IRC-Library, Information Resource Center der Jacobs University Bremen
Granting Institution:Jacobs Univ.
Author:Tomasz Łuczyński
Referee:Andreas Birk, Kaustubh Pathak, Nuno Gracias
Advisor:Andreas Birk
Persistent Identifier (URN):urn:nbn:de:gbv:579-opus-1007798
Document Type:PhD Thesis
Date of Successful Oral Defense:2017/12/08
Date of First Publication:2018/01/19
Academic Department:Computer Science & Electrical Engineering
PhD Degree:Computer Science
Focus Area:Mobility
Call No:Thesis 2017/38

$Rev: 13581 $