Numerical and Experimental Study of Wave Propagation in Water-Saturated Granular Media Using Effective Method Theories and a Full-Wave Numerical Simulation

Abstract

In this article, the detection of an object buried in marine sediments is investigated. Using the results from a series of experiments realized in a tank filled with water and calibrated glass beads, we evaluate the performances for a wide range of the value of the ratio kd of the grain size to the wavelength (k is the wave number and d is the grain diameter) of three types of prediction tools. The first two prediction tools are based on the definition of an equivalent model. The first tool is based on the well-known Biot-Stoll theory (BM model) while the second tool uses nonperiodic homogenization to define effective velocity and anisotropy maps representing the medium (HM model). The last prediction tool implements a timedomain full-wave numerical method, which takes into account each grain separately (GM model). It is shown that a good agreement between experiments and numerical simulations can be achieved using the BM model for the low kd regime and the HM model for high kd regime.

An efficient coral survey method based on a large-scale 3-D structure model obtained by Speedy Sea Scanner and U-Net segmentation

“…This precise seabed topography will contribute not only to coral surveys but also to other ecological, engineering and geographical studies, e.g., high-resolution advection modeling and structural calculations of natural reefs [39][40][41] . The survey efficiency of 12,146 m 2 /h achieved in this study is higher than the 7,000 m 2 /h of the previous study 26 , because six cameras were used in this case compared to five in the previous one due to battery problems. In addition, the water transparency was better than before (see the supplementary Fig.…”

“…Through recent technology innovation, a more efficient image collection system, namely the "Speedy Sea Scanner (SSS)" (Fig. 1), which is a towed optical camera array that has succeeded in making a large-scale and high-resolution 2-D image (orthophoto) of the seafloor around the Kujuku islands in 2017 26 . When that imaging was collected, the surveying efficiency of the SSS was approximately 7,000 m 2 /h.…”

“…Thus, the surveying efficiency of the SSS is a dramatic improvement over these earlier methods and we can now obtain a large number of images with greater ease than before. In addition, precise depth information on the seafloor can be obtained from a 3-D structure model derived from part of the survey area 26 . However, a large-scale 3-D structure model of an entire survey area has not been generated and the accuracy of the seafloor's depth distribution has yet to be evaluated.…”

“…However, a large-scale 3-D structure model of an entire survey area has not been generated and the accuracy of the seafloor's depth distribution has yet to be evaluated. In conjunction with the SSS technology's development, to reduce the time required for the analysis of huge quantities of data, an automatic coral coverage estimation method that makes use of conventional image segmentation approaches based on pixelwise CNN and bag-of-visual-words (BoVW) was proposed and the performances were compared 26 . In the comparison, the accuracy of pixelwise CNN was found to be better than that of BoVW.…”

“…In addition, we propose another segmentation method based on U-Net 28 , (which is often used in medical applications) 29 , and perform the coral cover estimation using the large-scale 2-D image (orthophoto) converted from the 3-D structure model. The computational cost of the U-Net-based segmentation method is much smaller than that of the pixelwise CNN-based one 26 . The prediction time of U-Net is about 1/1,000 for pixelwise CNN (See the Results section for the details).…”