Adaptive Seabed Characterization With Hierarchical Bayesian Modeling of SAS Imagery

Grand Challenges in Acoustic Remote Sensing: Discoveries to Support a Better Understanding of Our Changing Planet

“…These active systems have filled this sensor void in helping us understand underwater systems where EM sensors have difficulty reaching. Sonar (Sound Navigation and Ranging) systems primarily utilize frequencies ranging from 1 kHz to several hundreds of kHz (Lurton, 2010), and the design and engineering of these sensors has advanced and diversified for use in a wide range of application: Singlebeam echosounders (SBES) and multibeam echosounders (MBES) are used to map seafloor bathymetry for nautical charting (e.g., Mayer, 2006;Lurton, 2010;Brown et al, 2011;Mayer et al, 2018); MBES, sidescan sonar (SSS) and lower frequency sub bottom profilers (SBP) are used to study the morphology and geology of the seabed (e.g., Piper et al, 1999;Polyak et al, 2001;Collier and Brown, 2005;Wilson et al, 2007;Lecours et al, 2016); SBES, MBES and acoustic telemetry are used for fisheries applications to map fish biomass and movement (e.g., Mayer, 2002;Foote, 2009;Colbo et al, 2014;Crossin et al, 2017;Muñoz et al, 2020) or to map benthic ecosystems (e.g., Brown et al, 2011;Micallef et al, 2012;Ierodiaconou et al, 2018;Lacharite et al, 2018;Brown et al, 2019;Wilson et al, 2021); Synthetic aperture sonars (SAS) are primarily used for defense applications (e.g., Hayes and Gough, 2009;Myers and Fawcett, 2010) with other applications such as benthic habitat or substrate mapping recently emerging (Brandes and Ballard, 2019;Thorsnes et al, 2019); Acoustic Doppler Current Profilers (ADCP) are used to investigate physical oceanographic phenomena including current speed, direction and transport of biological or geological particles (e.g., Fielding et al, 2004;Gartner, 2004;Thomson et al, 2012). At lower frequencies (<1,000 Hz), seismic systems have been developed for remote characterization of the deep seafloor subsurface (e.g.,…”

An SVM-Based AdaBoost Cascade Classifier for Sonar Image

“…The sonar image classification is one of the most important research topics with great significance. Up to now, related works have involved the tracking and protection of endangered aquatic organisms [1], classification and tracking of sea-surface obstacles [2][3], water environment sounding and modeling [4], seabed modeling and mapping [5][6], salvage and rescue [7], submarine pipeline detection [8][9][10][11], submarine target location and identification [12][13][14], submarine environment description [15], and clinical Medical disease detection [16]. Generally, the color features of the images obtained by underwater equipment are usually lost, and the scattering of light and non-target also cause disturbance.…”