1. Antarctic krill (Euphausia superba) form aggregations known as swarms that vary greatly in size and density. Six acoustic surveys were conducted as part of multidisciplinary studies at two study sites, the western and eastern core boxes (WCB and ECB), during the 1997, 1998 and 1999 austral summers, at South Georgia. A quantitative, automated, image processing algorithm was used to identify swarms, and calculate swarm descriptors, or metrics. In contrast to acoustic surveys of aggregations of other pelagic species, a strong correlation (r = 0.88, p = 0.02, 95% C.I.= 0.24 to 0.99) between the number of krill swarms and the mean areal krill density [rho.hat] was found. Multivariate analysis was used to partition swarms into three types, based on contrasting morphological and internal krill density parameters. Swarm types were distributed differently between inter-surveys and between on and off-shelf regions. This swarm type variation has implications for krill predators, by causing spatial heterogeneity in swarm detectability, suggesting that for optimal foraging to occur, predators must engage in some sort of adaptive foraging strategy. 2. Krill predator-prey interactions were found to occur at multiple spatial and temporal scales, in a nested, or hierarchical structure. At the largest inter-survey scale, an index of variability, I, was developed to compare variation in survey-scale predator sightings, sea temperature and [rho.hat]. Using I and a two-way ANOVA, core box, rather than year, was found to be a more important factor in determining species distribution. The absence of Blue-petrels (Halobaena caerulea) and the elevated number of Antarctic fur seals (Arctocephalus gazella) suggest that 1998 was a characterised by colder than average water surrounding South Georgia, and a high [rho.hat] in the ECB. At the smaller, intra-survey scales (<80 km, <5 day), the characteristic scale (distances in which predator group size, or krill density were similar, L_s) were determined. For krill and predators L_s varied by survey and the L_s of krill also varied by depth within a survey. Overlap in L_s were stronger between predator species than between a predator species and krill, indicating predators were taking foraging cues from the activity of predators, rather than from the underlying krill distribution. No relationship was found between swarm characteristics and predator activity, suggesting either there is no relationship between krill swarms and predators, or that the predator and acoustic observation techniques may not be appropriate to detect such a relationship. 3. To overcome the 2-D sampling limitations of conventional echosounders, a multibeam echosounder (MBE) observed entire swarms in three-dimensions. Swarms found in the nearshore environment of Livingston Island situated in the South Shetland Islands, exhibited only a narrow range of surface area to volume ratios or roughnesses (R = 3.3, CV = 0.23), suggesting that krill adopt a consistent group behaviour to maintain swarm shape. Generalized additive models (GAM) suggested that the presence of air-breathing predators influenced the shape of a krill swarm (R decreased in the presence of predators: the swarm became more spherical). A 2D distance sampling framework was used to estimate the abundance, N, and associated variance of krill swarms. This technique took into account angular and range detectability (half-normal, [sigma_r.hat] = 365.00 m, CV = 0.16) and determined the vertical distribution of krill swarms to be best approximated by a beta-distribution ([alpha.hat] = 2.62, [CV.hat] = 0.19; [beta.hat] = 2.41, [CV.hat] = 0.15), giving the abundance of swarms in survey region as [N.hat] = 5,062 ([CV.hat] = 0.35). This research represents a substantial contribution to developing estimation of pelagic biomass using MBEs. 4. When using a single- or split-beam missing pings occur when the transmit or receive cycles are interrupted, often by aeration of the water column, under the echosounder transducer during rough weather. A thin-plate regression spline based approach was used to model the missing krill data, with knots chosen using a branch and bound algorithm. This method performs well for acoustic observations of krill swarms where data are tightly clustered and change rapidly. For these data the technique outperformed the standard MGCV GAM, and the technique is applicable for estimating acoustically derived biomass from line transect surveys.

Orientador: Mauricio Galo / Resumo: Um Levantamento Hidrográfico (LH) tem como principal meta a obtenção de dados para a edição e atualização de documentos náuticos, estes, voltados à segurança das atividades de navegação. Objetivando padronizar parâmetros de incerteza das cartas náuticas, a Organização Hidrográfica Internacional (OHI) define níveis mínimos de confiança para diferentes ordens. A sugestão dessas especificações foi internalizada pela Marinha do Brasil, responsável pela produção das cartas náuticas brasileiras, na NORMAM-25. Um desses parâmetros é a Incerteza Vertical Total máxima permitida, um indicador de qualidade da medição da profundidade. A informação de profundidade influencia no calado máximo permitido a uma embarcação para transitar em uma região com segurança, o que pode impactar inclusive nas limitações de transações comerciais em terminais portuários, uma vez que as profundidades estimadas com acurácia potencializam os parâmetros de operação dos portos. Por se tratar de um ambiente dinâmico, seja por ação da própria natureza ou devido a atividades antrópicas, a atualização de uma carta náutica deve ser uma preocupação constante. Como complemento à tradicional técnica de levantamento por meio de um ecobatímetro acoplado a embarcações, há a opção de se realizar um LH com o emprego da tecnologia LiDAR (Light Detection And Ranging) a partir de aeronaves, por meio de um aerolevantamento batimétrico por LiDAR (ALB – Airborne LASER Bathymetry), que operam com pulsos LASER na região verde do e... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: A Hydrographic Survey (HS) has as main goal to obtain data for editing and updating nautical documents, these, focused on the safety of navigation. In order to establish a standard of uncertainty parameters for nautical charts, the International Hydrographic Organization (IHO) defines minimum levels of confidence for different orders. The suggestion of these specifications was acknowledged by the Brazilian Navy, institution responsible to produce Brazilian nautical charts, as described in NORMAM-25. One such parameter is the maximum allowed Total Vertical Uncertainty, a quality indicator of the depth measurement. Depth information influences the maximum operational draft for a vessel to safely travel in a region, causing impact on port operations and limiting the commercial transactions. Accurately estimated depths enhance the operational parameters of the ports. Due to the aim of representing a dynamic environment, whether as consequence of the action of nature itself or because of anthropic activities, updating a nautical chart must be a constant concern. As a complement to the traditional survey technique conducted with a boat-coupled echosounder, there is the option of performing a HS using LiDAR (Light Detection And Ranging) technology from aircraft, through LiDAR aerial bathymetry (ALB - Airborne LASER Bathymetry), which operate with LASER pulses in the green region of the electromagnetic spectrum. Considering these points, this work analyzed the differences between the... (Complete abstract click electronic access below) / Mestre

L'objectif de la thèse est de développer une méthode de modélisation 3D des tunnels canaux en service, afin de les documenter de manière précise. Le levé des structures est effectué en dynamique depuis un bateau, avec un ensemble de caméras au-dessus de l'eau et un sonar pour la partie immergée. La construction du modèle 3D doit combiner des données acquises dans deux milieux différents, en l'absence de signal GPS pour la localisation du système de mesure. L'approche proposée exploite pleinement les capacités du calcul photogrammétrique, à la fois pour construire le modèle 3D de la voûte et pour estimer la trajectoire du bateau, laquelle permet de géoréférencer les profils de points sonar. L'application du procédé en vraie grandeur dans le tunnel de Niderviller a permis d'obtenir un premier modèle 3D. L'analyse de la méthode a mis en jeu une technologie innovante de mesure sonar statique, nécessitant le développement de traitements appropriés. Elle a permis de jauger les capacités du sonar à numériser les canaux, d'évaluer les trajectoires estimées par photogrammétrie et de comparer quantitativement le modèle obtenu à un modèle de référence préalablement construit. / The aim of this thesis is to develop a 3D modeling method for canal-tunnels in service, in order to document them accurately. The survey of the structures is carried out dynamically from a boat, with a set of cameras above the water and a sonar for the submerged part. The construction of the 3D model must combine data acquired in two different environments, in the absence of a GPS signal to locate the measurement system. The proposed approach makes full use of the capabilities of photogrammetric computation, both to build the 3D model of the vault and to estimate the trajectory of the boat, which allows georefencing of sonar point profiles. The application of the process to field test data acquired in the Niderviller tunnel allowed a first 3D model to be obtained. The analysis of the method involved an innovative static sonar measurement technology, requiring the development of appropriate treatments. It enabled to assess the capacity of the sonar for canal surveying, to evaluate the trajectories estimated using photogrammetry and to compare quantitatively the obtained model with a previously constructed reference mode

Using a towed underwater video camera system, benthic habitats were classified along transects in a popular offshore fishing area on the West Florida Shelf (WFS) known as “The Elbow.” Additionally, high resolution multibeam bathymetry and co-registered backscatter data were collected for the entire study area. Using these data, full coverage geologic and biotic habitat maps were developed using both unsupervised and supervised statistical classification methodologies. The unsupervised methodology used was k-means clustering, and the supervised methodology used a random forest algorithm. The two methods produced broadly similar results; however, the supervised methodology outperformed the unsupervised methodology. The results of the supervised classification demonstrated “substantial agreement” (κ>0.6) between observations and predictions for both geologic and biotic habitat, while the results of the unsupervised classification demonstrated “moderate agreement” (κ>0.4) between observations and predictions for both geologic and biotic habitat. Comparisons were made with the previously existing map for this area created by Florida Fish and Wildlife Conservation Commission’s Fish and Wildlife Research Institute (FWC-FWRI). Some features are distinguishable in both maps, but the FWC-FWRI map shows a greater extent of low relief hard bottom features than was predicted in our habitat maps. The areas predicted as low relief hard-bottom by FWC-FWRI often coincide with areas of higher uncertainty in the supervised map of geologic habitat from this study, but even when compared with ground-truth points from the towed video rather than predictions, the low relief hard bottom in FWC-FWRI’s map still corresponds to what was identified as sand in the video 73% of the time. The higher uncertainty might be a result of the presence of mixed habitats, differing morphology of hard-bottom, or the presence of sand intermixed with gravel or debris. More ground-truth samples should be taken in these areas to increase the confidence of these classifications and resolve discrepancies between the two maps. Data from the towed video system were also used to assess differences in fish communities among habitat types and to calculate habitat-specific densities for each taxa. Fish communities were found to significantly differ between soft and hard bottom habitats as well as among the hard-bottom habitats with different vertical relief (flat hard-bottom vs more steeply sloping areas). Additionally, significant differences were found between the fish communities in habitats with attached fauna such as sponges and gorgonians, and areas without attached fauna; however, attached fauna require rock to attach to and the rock habitats rarely lacked attached fauna, so this difference may just reflect the difference between fish communities in sand and rock habitats without the consideration of vertical relief. Moreover, the species driving the differences in the fish communities were identified. Fish were more likely to be present and assemblages were more species rich in more complex habitats (rockier, higher relief, presence of attached fauna). Habitat specific densities were calculated for each species, and general trends are discussed. Lastly the habitat-specific densities were extrapolated to the total area of habitat type (sand vs rock) as predicted by the supervised geologic habitat map. There is predicted to be approximately 111,000 fish (95% CI [67015, 169405]) within the study area based on this method, with ~47,000 (~43%) predicted to be within the sand habitat and ~64,000 (~57%) in the rock habitat. This demonstrates the potential of offshore rocky reefs as “critical habitats” for demersal fish in the offshore environment as rock accounts for just 4% of the study area but is expected to contain over half of the total abundance. The value of sand habitats is also shown, as due to their large area they are able to contribute substantially to the total number of fish despite sustaining comparatively low densities.

The former Viking settlement Birka is located on Björkö Island in Lake Mälaren, the third largest lake in Sweden. Birka is a well-known archeological site that onshore has been carefully examined. The lake floor of the waters surrounding the island has been less investigated but has a great potential to host not yet discovered archeological objects from this former hub for seafarers. Therefore, a geophysical survey including multibeam sonar mapping and subbottom profiling was carried out mainly along the shores of western Björkö Island. Processing and analysis of these collected data form the basis of this thesis. The main aims of this study are to produce a suite of geological maps and stratigraphic profiles that are used to geologically interpret the uppermost sediment stratigraphy and the bathymetry of the area. In addition, the processed data are investigated for archeological objects. The result shows that the acoustic records of the sediment stratigraphy reaches back to glacial clay formed as a consequence of the retreat of the Late Weichselian ice sheet, and that the uppermost sediment units probably are from the time after the isolation of Lake Mälaren from the Baltic Sea. The bathymetry and backscatter results have revealed that this relatively shallow study area contains several objects that potentially could be of interest from an archeological point of view. These objects include several unidentified objects in the Björkö strait and two unregistered shipwrecks where ground truthing data need to be collected to determine their respective origin.

La mesure de Matières En Suspension (MES) est cruciale autant pour comprendre les transferts sédimentaires que pour les études des écosystèmes marins. Elles sont classiquement mesurées ponctuellement par des prélèvements d’eau in situ, ou à partir des propriétés optiques de l’eau. Mais depuis plusieurs décennies, les appareils acoustiques, ont montré leur capacité à mesurer ces MES sur de plus grands volumes. Ces mesures, en particulier celles de la concentration, s’appuient sur les propriétés de rétrodiffusion des particules. Mais ces mesures demeurent peu représentatives dans la mesure où le contenu en MES dans la colonne d’eau varie à différentes échelles de temps et d’espace. Dans un premier temps, ces travaux de thèse visent à montrer à travers des mesures in situ réalisées dans l’estuaire de l’Aulne qu’il est possible d’étendre spatialement la mesure par inversion des données issues de sondeurs multifaisceaux (SMF). Un modèle de rétrodiffusion adapté à la suspension d’intérêt, constituée dans cette étude d’agrégats estuariens, est d’abord élaboré. Puis grâce à des observations issues d’un profileur multifréquences, la concentration massique en MES, distribuée par classes de tailles, a pu être déterminée par la résolution d’un problème inverse. Ces informations permettent de déterminer les rayons équivalents de la suspension, grâce auxquels les données issues du SMF, au préalable calibrées par une méthode innovante, peuvent être inversées de manière directe. Une étude des incertitudes attachées aux valeurs de concentrations estimées est par la suite proposée afin de qualifier la pertinence des résultats. Puis dans un second temps, les conditions nécessaires à l’établissement d’un protocole de mesure des MES par système multi-capteur sont identifiées. Ce dernier permet la caractérisation en continu des MES à différentes échelles de temps et d’espace, en exploitant la complémentarité des mesures issues des différents instruments. / Measuring Suspended Particulate Matter (SPM) is essential to better understand sediment transport and marine ecosystems. SPM is traditionnaly estimated through in situ water samples analysis, or based on the optical properties of water. Yet for several decades, acoustical devices have shown their capability to measure SPM on larger volumes. These measurements (especially in terms of SPM concentration) are based on the backscattering properties of the particles. However, these measurements remain limited, since the SPM content in the water column is subjected to variations on both spatial and temporal scales.As a first step, this work aims at showing that it is possible to increase the degree of spatialization of the SPM measurements by inverting MultiBeam EchoSounder (MBES) data, through in situ measurements acquired in the Aulne macrotidal estuary. A backscattering model was first designed to describe the backscattering properties of the suspension of interest, consisting in this study in estuarine agregates. Then, thanks to multifrequency observations, the SPM mass concentration sorted by size classes was retrieved through the the resolution of the inverse problem. This kind of information allows to determine the equivalent spherical radius of the the whole suspension, through which the MBES data, calibrated beforehand using an original method, can be directly inverted. Subsequently, a study of the uncertainties attached to the final concentration estimate is proposed in order to qualify the relevance of the results.In a second step, the necessary conditions for establishing a measurement protocol of the SPM are identified. The latter allows continuous characterization of SPM at different spatial and temporal scales, by exploiting the complementarity of the measures delivered by different instruments.

Sjöfartsverket har i sitt arkiv en stor mängd backscatterdata, insamlat med multibeamekolod, som ännu inte har använts till sin fulla potential. Backscatterdata innehåller information om den reflekterade signalens styrka, även kallad amplitud. Stora mängder backscatterdata kan användas för att beskriva den akustiska bottenreflektionen. Syftet med denna undersökning är att beskriva hur variationen för amplituden varierar beroende på vilken bottentyp den reflekteras ifrån. En metod för att skapa rasterdataset med bottenhårdhet och bottentyp baserat på amplituddata ska utvecklas. Resultaten från denna metod ska sedan jämföras med kartdata från Sveriges Geologiska Undersökning (SGU). Totalt användes cirka 45 miljoner bottenpunkter i studieområdet. Varje punkt innehåller information om amplitud som systemet har registrerat från det reflekterade ekot. Dessa data behövde genomgå databehandlingar, bl.a. en vinkelkorrigering som ger ett mer trovärdigt värde av amplitud. Med hjälp av befintlig information om studieområdets sjöbotten i form av en maringeologisk karta från SGU, kunde amplitud från ett antal specifika uppskattade bottentyper studeras direkt. Resultatet uppvisar stora skillnader i amplitudens variationer. Specifika medelvärden och standardavvikelser kan urskiljas beroende av vilken specifik uppskattad bottentyp som studerades. ”Mjuk lera” gav en svagare signal med relativt låg standardavvikelse. ”Häll” och ”sten och block” reflekterade en liknade men starkare signal. Amplitudata från backscatter-informationen i hela datamängden utnyttjades för att skapa raster vars syfte var att beskriva den uppskattade bottenhårdheten. Olika raster skapades med olika parametrar beroende på ändamål. Gemensamt för alla skapade raster är att de är uppbyggda med metoden ”flytande beräkning” som möjliggjorde en mer utjämning. Resultatet av medelvärde och standardavvikelse från varje enskild bottentyp utnyttjades för att utföra en klassning av bottentyper på ett skapat raster lämpad för just bottentypsklassificering. För att få ett mer noggrannare medelvärde och standardavvikelse studerades ett 68 % konfidensintervall för de olika bottentyperna. De bottentyper som valdes för klassningen var ”mjuk lera”, ”sand, grus och sten”, ”häll”, ”sten och block” och även ”lägre amplituder”. ”Häll” och ”sten och block” klassades samma eftersom deras fysikaliska egenskaper gör att deras värden ligger nära varandra vilket gjorde det svårt att urskilja dem.”Lägre amplituder” utnyttjades för att identifiera områden som har lägre reflektionsförmåga än mjuk lera. Vilken bottentyp det är kan endast provtagning ge svar på. Med hjälp av tolkning av skapade raster och den maringeologiska kartan så korrigerades intervallen och användes som klassning. Resultatet från klassningen visar tydligt att områden kan urskiljas i kartbilden. Majoriteten av klassningarna resulterade i typen mjuk lera. En jämförelse av klassningen med den maringeologiska kartan visar att stora skillnader finns mellan dem. Mjuk lera gav en överensstämmelse på 86 %, sand, grus och sten 30 % och häll, sten och block 52,5 %, vilket gav en total överenstämmelse på 56,2 %. Jämförelse utfördes även med 9 provtagningspunkter som fanns tillgängliga i området. Det visade en total överenstämmelse på 89 %. Undersökningen visar att amplitud från havsbottnen korrelerar med bottentypen det är. Noterbart är att metoden för bottentypsklassificering som utvecklats i denna studie inte har kunnat kvalitetsgranskas med ett trovärdigt resultat, p.g.a. av statistiskt för få provtagningspunkter att jämföra mot. Studien visar dock att med mer data och noggrannare referensdata kan en mer automatisk klassningsmetod utvecklas. / The Swedish Maritime Administration (Sjöfartsverket) has a large amount of backscatter data collected with multibeam echosounder in their archive that has not been fully used despite its great potential. Backscatter data contains information about the strength of the reflected signal, often called amplitude strength. Large amounts of backscatter data could be used to describe the acoustic bottom reflection. The purpose of this study is to describe how the variation of the amplitude strength varies dependent on which estimated bottom types the data reflects from. Also a method will be produced which purpose is to create gridded dataset of estimated bottom hardness and bottom type based on amplitude data and compare this method with official data from the Geological Survey of Sweden (SGU).A total number of 45 million depths (data points) were used in the study area. Every data point contains information about the amplitude strength that the system has recovered from the reflective echo. This data needed to be preprocessed, including an angle correction that produces a more reliable value of the amplitude strength. With existing information about the bottom from the study area, in this case a marine geological map from SGU, the amplitude from some estimated specified bottom types could be studied. The result shows differences in their variation. Specific values of mean and standard deviation could be distinguished by which estimated specific bottom types that were studied.The amplitude strength from the backscatter information of the complete data set was used to create a raster that describes the estimated bottom-hardness. Different raster were created with various parameters dependent on the purpose. All of the created raster data had in common that it was created using a technique called “flow calculation” which result in more equalization.The mean and standard deviation for every individual estimated bottom type were used to create interval for classification of the bottom types. To achieve a more accurate estimation of the mean and standard deviation for the bottom types, a 68 % confidence interval were used. The classes that were chosen for classification was “soft clay”, “sand, gravel and stone”, “solid rock”, “stone and block” and “lower amplitudes”. “Solid rock” and “stone and block” were combined in the same class because of their similar physical properties. “Lower amplitudes” were chosen in order to indicate areas where the amplitude strength from the reflective echo was lower than “soft clay”.The result of the intervals was adjusted by an examination of the raster data and the marine geological map and was then used for classification.The result from the classification shows that areas of different bottom types could be distinguished in the map. The majority of the classification was of the type “soft clay”. A comparison between the classification and the marine geological map showed some differences. “Soft clay” matched with 86 %, “sand, gravel and stone” 30 %, “solid rock, stone and block” 52,5 % and the total matched with 56,2 %. Comparisons between 9 samplings in the area were made. The result shows that the classification-accuracy is 89 %.The study shows that the amplitude strength correlates to the bottom type. Note that too few samplings for bottom classification were used in the study and thus the results are not fully reliable. The study, however, shows that with larger amount of data and more accurate reference data a better automatic classification method could be developed.

Cette thèse, co-dirigée par Jean-Marc Boucher et Ronan Fablet (IMT Atlantique) et co-encadrée par Didier Charlot (iXBlue), Gilles Le Chenadec et Michel Legris (ENSTA Bretagne), a été réalisée dans le cadre d'une convention CIFRE au sein de la société iXBlue. iXblue développe et commercialise un sondeur multifaisceaux (MBES) SEAPIX principalement dédié au marché de la pêche. Ce système a été développé pour offrir le meilleur compromis entre performances de détection et son coût de revient. Outre les caractéristiques classiques d'un MBES, il propose la particularité unique de pouvoir insonifier des fauchées différentes sous le navire par dépointage électronique du faisceau d'émission de bâbord à tribord et d'avant en arrière. Le travail de thèse a pour objectif d'étudier l'apport de ces nouvelles capacités multi-fauchées dans l'analyse et la classification des fonds marins. La première partie du travail a consisté à réaliser une analyse détaillée de la chaîne de mesure. Cette étude a permis d'évaluer la consistance des niveaux de rétrodiffusion entre les différents modes d'insonification. La deuxième partie s'est intéressée à la recherche des caractéristiques discriminantes du signal rétrodiffusé en tenant compte de la géométrie d'acquisition de chaque mode d'insonification. La dernière étape du travail a porté sur des méthodes de fusion des données acquises. Cette étude s'est réalisée en deux approches; la première considère des données venant du même mode d'insonification (intra-mode) et la seconde venant de modes différents (inter-mode), pour la cartographie des fonds marins. Les résultats expérimentaux obtenus mettent en évidence l'intérêt de la chaîne de traitement proposée et d'une architecture multi-mode sur les jeux de données réelles traitées. / This thesis, co-directed by Jean-Marc Boucher and Ronan Fablet (IMT Atlantique) and co-supervised by Didier Charlot (iXBlue), Gilles Le Chenadec and Michel Legris (ENSTA Bretagne), was realized in the context of a convention CIFRE with the company iXBlue.iXblue develops and commercializes a multibeam echosounder (MBES) SEAPIX primarily dedicated to the fishery market. The system is optimized to offer the best compromise between performances capabilities and cost. In addition to the classical characteristics of an MBES, it offers the unique feature of scanning the seafloor (and the water column volume) by electronical beamform multiple the emission swaths from port to starboard, as well as from forward to backward. The objective of the thesis is to study the contribution of these new multi-swath capacities in the analysis and classification of the seafloor.The first part of the work consisted in carrying out a detailed analysis of the measurement chain. This study evaluated the consistency in acquiring the backscattering strength from different insonification modes. The second part investigated the discriminant characteristics of the backscattered signal while taking into account the acquisition geometry of each insonification mode. The last stage of the work involved to methods of fusing the acquired data. This study was carried out in two approaches; the first considers data from the same insonification mode (intra-mode) and the second from different modes (inter-mode), for the seafloor classification. The obtained experimental results highlight the interest of the proposed processing chain and a multi-mode architecture on the real datasets.

Underwater navigation is a key technology for exploring the oceans and exploiting their resources. For autonomous underwater vehicles (AUVs) to explore the marine environment efficiently and securely, underwater simultaneous localization and mapping (SLAM) systems are often indispensable due to the lack of the global positioning system (GPS). In an underwater SLAM system, an AUV maps its surroundings and estimates its own pose at the same time. The pose of the AUV can be predicted by dead reckoning, but navigation errors accumulate over time. Therefore, sensors are needed to calibrate the state of the AUV. Among various sensors, the multibeam echosounder (MBES) is one of the most popular ones for underwater SLAM since it can acquire bathymetric point clouds with depth information of the surroundings. However, there are difficulties in data association for seabeds without distinct landmarks. Previous studies have focused more on traditional computer vision methods, which have limited performance on bathymetric data. In this thesis, a novel method based on deep learning is proposed to facilitate underwater perception. We conduct two experiments on place recognition and point cloud registration using data collected during a survey. The results show that, compared with the traditional methods, the proposed neural network is able to detect loop closures and register point clouds more efficiently. This work provides a better data association solution for designing underwater SLAM systems. / Undervattensnavigering är en viktig teknik för att utforska haven och utnyttja deras resurser. För att autonoma undervattensfordon (AUV) ska kunna utforska havsmiljön effektivt och säkert är underwater simultaneous localization and mapping (SLAM) system ofta oumbärliga på grund av bristen av det globala positioneringssystemet (GPS). I ett undervattens SLAM-system kartlägger ett AUV sin omgivning och uppskattar samtidigt sin egen position. AUV:s position kan förutsägas med hjälp av dödräkning, men navigeringsfel ackumuleras med tiden. Därför behövs sensorer för att kalibrera AUV:s tillstånd. Bland olika sensorer är multibeam ekolod (MBES) en av de mest populära för undervattens-SLAM eftersom den kan samla in batymetriska punktmoln med djupinformation om omgivningen. Det finns dock svårigheter med dataassociation för havsbottnar utan tydliga landmärken. Tidigare studier har fokuserat mer på traditionella datorvisionsmetoder som har begränsad prestanda för batymetriska data. I den här avhandlingen föreslås en ny metod baserad på djup inlärning för att underlätta undervattensuppfattning. Vi genomför två experiment på punktmolnregistrering med hjälp av data som samlats in under en undersökning. Resultaten visar att jämfört med de traditionella metoderna kan det föreslagna neurala nätverket upptäcka slingförslutningar och registrera punktmoln mer effektivt. Detta arbete ger en bättre lösning för dataassociation för utformning av undervattens SLAM-system.

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