Joint underwater mapping with acoustic and optical Iimages
This thesis is developed within the context of the IURBI project [1], which seeks to develop an intelligent AUV capable of real-time seafloor analysis and adaptive mission planning (Figure 1.1). A fundamental prerequisite for such autonomous capabilities is the ability to robustly align and fuse sen...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Fecha de publicación: | 2025 |
| País: | España |
| Institución: | Varias* (Consorci de Biblioteques Universitáries de Catalunya, Centre de Serveis Científics i Acadèmics de Catalunya) |
| Repositorio: | Recercat. Dipósit de la Recerca de Catalunya |
| OAI Identifier: | oai:recercat.cat:10256/28379 |
| Acceso en línea: | http://hdl.handle.net/10256/28379 https://hdl.handle.net/10256/28379 |
| Access Level: | acceso abierto |
| Palabra clave: | Autonomous Underwater Vehicles Autonomous Underwater Vehicles -- Navigation systems Vehicles submergibles autònoms -- Sistemes de navigació Digital mapping Cartografia digital SLAM Sonar Sonar (Navegació) Algorismes Algorithms |
| Sumario: | This thesis is developed within the context of the IURBI project [1], which seeks to develop an intelligent AUV capable of real-time seafloor analysis and adaptive mission planning (Figure 1.1). A fundamental prerequisite for such autonomous capabilities is the ability to robustly align and fuse sensor data from multiple sources and surveys into a single, coherent model. This thesis addresses that foundational challenge by developing a comprehensive offline framework for multi-session, multimodal map alignment. The primary objectives of this thesis are to: – Develop a robust and flexible framework for the alignment and integration of side-scan sonar and optical imagery acquired in single or multiple sessions by AUVs, towfish, or ROVs. – Formulate and implement a factor graph optimization approach to jointly re fine vehicle trajectories and sensor alignments across multiple sessions and modalities, accommodating the inherent uncertainties in underwater navigation. – Evaluate the performance of the proposed methodology using real-world under water datasets, assessing its accuracy, robustness, and practical applicability. The scope of this work encompasses the offline processing and alignment of pre viously collected side-scan sonar and optical image datasets. While initial navigation data from the AUV/ROV is assumed to be available, this work specifically focuses on refining these initial pose estimates to achieve precise multimodal and multi-session co-registration. |
|---|