OceanRep
Visual 3D Reconstruction in the Deep Sea : from Geometric and Radiometric Perspectives.
Song, Yifan (2024) Visual 3D Reconstruction in the Deep Sea : from Geometric and Radiometric Perspectives. (PhD/ Doctoral thesis), Christian-Albrechts-Universität, Kiel, Germany, X, 147 pp.
Preview |
Text
Dissertation_Yifan_Song_2024.pdf - Published Version Available under License Creative Commons: Attribution 4.0. Download (168MB) | Preview |
Abstract
Underwater vision systems have become popular in ocean research. However, imaging in the deep sea poses significant challenges. Firstly, cameras enclosed in waterproof housings must withstand extreme water pressure, while light undergoes refraction as it traverses multiple media layers. Secondly, light absorption and scattering in water degrade image quality. Additionally, artificial illumination in the deep sea creates unique visual effects. All these factors collectively contribute to the complexity and difficulty of visual 3D reconstruction in the deep sea.
This dissertation provides comprehensive solutions to the challenges encountered in deep sea visual 3D reconstruction, encompassing both hardware and software components. To address geometric refraction effects, it thoroughly analyzes dome refractive geometry and proposes mechanical adjustments for centering cameras inside domes to mitigate refraction effects. Additionally, recognizing the wide use of flat port interfaces, the dissertation explores effective methods to correct flat refraction effects for underwater photogrammetry.
Furthermore, the dissertation delves into radiometric issues arising from light absorption and scattering in water, especially concerning artificial illumination in deep-sea imaging. It explores physical models of underwater image formation and develops a general model suitable for restoring various types of underwater image, especially in environments with complex artificial illumination. This model enables the simultaneous removal of artificial lighting patterns and the restoration of true colors.
Moreover, the dissertation demonstrates several successful applications of the techniques developed within its scope. It showcases their efficacy in various camera system designs, underwater 3D reconstruction, and improving complex illumination configurations within imaging systems.
Document Type: | Thesis (PhD/ Doctoral thesis) |
---|---|
Thesis Advisor: | Köser, Kevin |
Keywords: | Underwater Photogrammetry; Underwater Image Formation; Refractive Geometry; Underwater Image Restoration; Seafloor Mapping |
Research affiliation: | OceanRep > GEOMAR > FB2 Marine Biogeochemistry > FB2-MG Marine Geosystems |
Main POF Topic: | PT6: Marine Life |
Date Deposited: | 25 Jun 2024 13:48 |
Last Modified: | 25 Jun 2024 13:48 |
URI: | https://oceanrep.geomar.de/id/eprint/60490 |
Actions (login required)
View Item |
Copyright 2023 | GEOMAR Helmholtz-Zentrum für Ozeanforschung Kiel | All rights reserved
Questions, comments and suggestions regarding the GEOMAR repository are welcomed
at bibliotheksleitung@geomar.de !