This study focuses on the potential of using a single beam echosounder as a tool for recognition and assessment of seafloor vegetation. Seafloor vegetation is plant benthos and occupies a large portion of the shallow coastal bottoms. It plays a key role in maintaining the ecological balance by influencing the marine and terrestrial worlds through interactions with its surrounding environment. Understanding of its existence on the seafloor is essential for environmental managers. / Due to the important role of seafloor vegetation to the environment, a detailed investigation of acoustic methods that can provide effective recognition and assessment of the seafloor vegetation by using available sonar systems is necessary. One of the frequently adopted approaches to the understanding of ocean environment is through the mapping of the seafloor. Available acoustic techniques vary in kinds and are used for different purposes. Because of the wide scope of available techniques and methods which can be employed in the field, this study has limited itself to sonar techniques of normal incidence configuration relative to seafloors in selected regions and for particular marine habitats. For this study, a single beam echosounder operating at two frequencies was employed. Integrated with the echosounder was a synchronized optical system. The synchronization mechanism between the acoustic and optical systems provided capabilities to have very accurate groundtruth recordings for the acoustic data, which were then utilized as a supervised training data set for the recognition of seafloor vegetation. / In this study, results acquired and conclusions made were all based on the comparison against the photographic recordings. The conclusion drawn from this investigation is only as accurate as within the selected habitat types and within very shallow water regions. / In order to complete this study, detailed studies of literature and deliberately designed field experiments were carried out. Acoustic data classified with the help of the synchronized optical system were investigated by several methods. Conventional methods such as statistics and multivariate analyses were examined. Conventional methods for the recognition of the collected data gave some useful results but were found to have limited capabilities. When seeking for more robust methods, an alternative approach, Genetic Programming (GP), was tested on the same data set for comparison. Ultimately, the investigation aims to understand potential methods which can be effective in differentiating the acoustic backscatter signals of the habitats observed and subsequently distinguishing between the habitats involved in this study.

Detailed information about the oceanic environment is essential for many applications in the field of marine geology, marine biology, coastal engineering, and marine operations. Especially, knowledge of the properties of the sediment body is often required. Acoustic remote sensing techniques have become highly attractive for classifying the sea bottom and for mapping the sediment properties, due to their high coverage capabilities and low costs compared to common sampling methods. In the last decades, a number of different acoustic devices and related techniques for analyzing their signals have evolved. Each sensor has its specific application due to limitations in the frequency range and resolution. In practice, often a single acoustic tool is chosen based on the current application, supported by other non-acoustic data where required. However, different acoustic remote sensing techniques can supplement each other, as shown in this thesis. Even more, a combination of complementary approaches can contribute to the proper understanding of sound propagation, which is essential when using sound for environmental classification purposes. This includes the knowledge of the relation between acoustics and sediment properties, the focus of this thesis. Providing a detailed three dimensional picture of the sea bottom sediments that allows for gaining maximum insight into this relation is aimed at.<p><p><p>Chapters 4 and 5 are adapted from published work, with permission:<p>DOI:10.1121/1.3569718 (link: http://asadl.org/jasa/resource/1/jasman/v129/i5/p2878_s1) and<p>DOI:10.1109/JOE.2010.2066711 (link: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?tp=&arnumber=5618582&queryText%3Dsiemes)<p>In reference to IEEE copyrighted material which is used with permission in this thesis, the IEEE does not endorse any of the Université libre de Bruxelles' products or services.<p> / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished