Active learning for detection of mine-like objects in side-scan sonar imagery

A data-adaptive algorithm is presented for the selection of the basis functions and training data used in classifier design with application to sensing mine-like targets with a side-scan sonar. Automatic detection of mine-like targets using side-scan sonar imagery is complicated by the variability of the target, clutter, and background signatures. Specifically, the strong dependence of the data on environmental conditions vitiates the assumption that one may perform a priori algorithm training using separate side-scan sonar data collected previously. In this paper, a novel active-learning algorithm is developed based on kernel classifiers with the goal of enhancing detection/classification of mines without requiring an a priori training set. It is assumed that divers and/or unmanned underwater vehicles (UUVs) may be used to determine the binary labels (target/clutter) of a small number of signatures from a given side-scan collection. These sets of signatures and associated labels are then used to train a kernel-based algorithm with which the remaining side-scan signatures are classified. Information-theoretic concepts are used to adaptively construct the form of the kernel classifier and to determine which signatures and associated labels would be most informative in the context of algorithm training. Using measured side-looking sonar data, the authors demonstrate that the number of signatures for which labels are required (via diver/UUV) is often small relative to the total number of potential targets in a given image. This procedure designs the detection/classification algorithm on the observed data itself without requiring a priori training data and also allows adaptation as environmental conditions change.     

浅水高分辨率测深侧扫声纳系统及其海上应用

介绍浅水高分辨率测深侧扫声纳系统(SBSSS)的工作原理和硬件结构,然后介绍系统进行海上井场区域调查的情况,最后给出系统的实时工作结果和水深数据后处理结果.并把系统获得的水深和地貌结果分别与EM3000D多波束测深系统的水深结果、Klein2000型侧扫声纳系统的地貌结果进行对比分析.比较结果表明,浅水高分辨率测深侧扫声纳系统与其他系统获得的结果一致,经数据后处理,测深精度达到IHO特级标准要求.     

Active exchange of water and nutrients between seawater and shallow pore water in intertidal sandflats

In order to determine the temporal and spatial variations of nutrient profiles in the shallow pore water columns (upper 30 cm depth) of intertidal sandflats, we measured the salinity and nutrient concentrations in pore water and seawater at various coastal environments along the southern coast of Korea. In the intertidal zone, salinity and nutrient concentrations in pore water showed marked vertical changes with depth, owing to the active exchange between the pore water and overlying seawater, while they are temporally more stable and vertically constant in the sublittoral zone. In some cases, the advective flow of fresh groundwater caused strong vertical gradients of salinity and nutrients in the upper 10 cm depth of surface sediments, indicating the active mixing of the fresher groundwater with overlying seawater. Such upper pore water column profiles clearly signified the temporal fluctuation of lower-salinity and higher-Si seawater intrusion into pore water in an intertidal sandflat near the mouth of an estuary. We also observed a semimonthly fluctuation of pore water nutrients due to spring-neap tide associated recirculation of seawater through the upper sediments. Our study shows that the exchange of water and nutrients between shallow pore water and overlying seawater is most active in the upper 20 cm layer of intertidal sandflats, due to physical forces such as tides, wave set-up, and density-thermal gradient.     

联合YOLOv4和迁移学习的侧扫声纳图像沉船检测

为解决侧扫声纳(SSS)图像沉船检测中样本不足、代表性弱等带来的检测精度低的问题,提出了一种联合YOLOv4和迁移学习的SSS图像沉船检测方法。首先,基于SSS成像机理实现了SSS沉船图像样本扩增,解决样本少而无法开展沉船检测模型构建的难题;然后,利用迁移学习,将公共数据集上学习到的权重和沉船通用性特征引入YOLOv4网络,构建高性能沉船检测模型。试验表明,构建的沉船检测模型取得了85.5%的类平均精度(mAP),将传统方法的检测精度提升了7.7%,在少样本情况下实现了沉船的高精度检测。     

Active impulsive echo discrimination in shallow water by mappingtarget physics-derived features to classifiers

One of the most difficult challenges in shallow-water active sonar processing is false-alarm rate reduction via active classification. In impulsive-echo-range processing, an additional challenge is dealing with stochastic impulsive source variability. The goal of active classification is to remove as much clutter as possible while maintaining an acceptable detection performance. Clutter in this context refers to any non-target, threshold-crossing cluster event. In this paper, we present a clutter-reduction algorithm using an integrated pattern-recognition paradigm that spans a wide spectrum of signal and image processing-target physics, exploration of projection spaces, feature optimization, and mapping the decision architecture to the underlying good-feature distribution. This approach is analogous to a classify-before-detect strategy that utilizes multiple informations to arrive at the detection decision. After a thorough algorithm evaluation with real active sonar data, we achieved over an order of magnitude performance improvement in clutter reduction with our methodology over that of the baseline processing     

High-frequency reverberation in shallow water

Monostatic reverberation measurements were collected in shallow water, over a coarse gravel and cobble bottom, 100 m deep, off the coast of Nova Scotia. Data were collected at frequencies of 21, 28, and 36 kHz using linear FM pulses of 2-kHz bandwidth and 0.160-s duration. An anchored, high-frequency active sonar array deployed at a depth of 42 m was used to collect the data. The reverberation measurements were compared with estimates computed with the NUWC generic sonar model (GSM). The data were reasonably well modeled for times greater than 0.2 s after pulse transmission by neglecting surface reverberation and using Lambert's rule for bottom backscattering with a scattering coefficient of -27 dB, independent of frequency. At all three frequencies, the data and model show a peak approximately 0.9 s after pulse transmission. This peak results from a focusing effect that the downward-refracting sound-speed profile has on the interaction of the rays with the bottom     

Propagation of shallow water waves in an open parabolic channel using the WKB perturbation technique

A singular perturbation analysis based on the WKB technique to study the hydrodynamic performance of periodic ocean waves that are incident on an open parabolic channel of constant depth is proposed. We derive a linear model to predict the propagation of the long ocean waves into the channel. In this manner, the spatial distribution for the surface elevation of the ocean waves inside the channel as a function of two dimensionless parameters, namely, a kinematical parameter, κ and a geometrical parameter ε, is governed by a second-order ordinary differential equation. The kinematical parameter κ denotes the ratio of the potential head, due to gravity, to the kinetic head of the ocean waves along the longitudinal axis of the parabolic channel. Meanwhile, ε is a dimensionless geometrical parameter that represents a characteristic ratio of the parabolic channel. Using matching conditions, simple expressions for the reflection and transmission coefficients are obtained.     

Numerical modeling of long waves in shallow water using Incremental Differential Quadrature Method

Incremental Differential Quadrature Method (IDQM) as a rapid and accurate method for numerical simulation of Nonlinear Shallow Water (NLSW) waves is employed. To the best of authors’ knowledge, this is the first endeavor to exploit DQM in coastal hydraulics. The one-dimensional NLSW equations and related boundary conditions are discretized in space and temporal directions by DQM rules and the resulting system of equations are used to compute the state variables in the entire computational domain. It was found that the splitting of total simulation time into a number of smaller time increments, could significantly enhance the performance of the proposed method. Furthermore, results of this study show two main advantages for IDQM compared with other conventional methods, namely; unconditional stability and minimal computational effort. Indeed, using IDQM, one can use a few grid points (in spatial or time direction) without imposing any stability condition on the time step to obtain an accurate convergent solution.     

Seismo-acoustic field statistics in shallow water

The spatial statistics of the acoustic field in shallow water are strongly affected by interfacial roughness and volume fluctuations in the water column or the seabed. These features scatter energy, reducing the coherence of the acoustic field. This paper introduces a consistent, mode-based modeling framework for ocean scattering. First, the rough surface scattering theory of Kuperman and Schmidt is reformulated in terms of normal modes, resulting in computation times which are reduced by several orders of magnitude. Next, a perturbation theory describing scattering from sound speed and density fluctuations in acoustic media is developed. The scattering theories are combined with KRAKEN, creating a unified normal mode code for wave theory modeling of shallow-water spatial statistics. The scattered field statistics are found to be a complicated function of scattering mechanism, scatterer statistics, and acoustic environment. Bottom properties, including elasticity, strongly influence the scattered field     

Wave-induced seabed response in shallow water

To simulate the wave-induced response of coupled pore fluids and a solid skeleton in shallow water, a set of solutions with different formulations (fully dynamic, partly dynamic, and quasi-static) corresponding to each soil behavior assumption is presented. To deal with Jacobian elliptic functions involved in the cnoidal theory, a Fourier series approximation is adopted for expanding the boundary conditions on the seabed surface. The parametric study indicates the significant effect of nonlinearity for shallow water wave, which also enhances the effect of soil characteristics. The investigation of the applicability of reduced formulations reveals the necessity of a partly or even fully dynamic formulation for the wave-induced seabed response problem in shallow water, especially for thickened seabed. The analysis of liquefaction in the seabed indicates that the maximum depth of liquefaction is shallower, and the width of liquefaction is broader under cnoidal wave loading. The present analytical model can provide more reasonable result for the wave-induced seabed response in the range of shallow water wave.     

Wave height distribution in shallow water

Probability distribution of shallow water wave heights, obtained from a pressure type recorder, are examined. It is tested with the theoretical distributions of (a) Rayleigh, (b) Weibull, (c) Gluhovski, (d) Ibrageemov and (e) Goda. The best fit is shown by the Gluhovski probability density function with a correlation coefficient greater than 0.8. The functions of Weibull, Ibrageemov and Goda fit only half of the tested cases. The role of wave steepness in the wave height distribution is found to be negligible.     

Numerical simulation of wave overtopping of coastal structures using the non-linear shallow water equations

A one-dimensional high-resolution finite volume model capable of simulating storm waves propagating in the coastal surf zone and overtopping a sea wall is presented. The model (AMAZON) is based on solving the non-linear shallow water (NLSW) equations. A modern upwind scheme of the Godunov-type using an HLL approximate Riemann solver is described which captures bore waves in both transcritical and supercritical flows. By employing a finite volume formulation, the method can be implemented on an irregular, structured, boundary-fitted computational mesh. The use of the NLSW equations to model wave overtopping is computationally efficient and practically flexible, though the detailed structure of wave breaking is of course ignored. It is shown that wave overtopping at a vertical wall may also be approximately modelled by representing the wall as a steep bed slope. The AMAZON model solutions have been compared with analytical solutions and laboratory data for wave overtopping at sloping and vertical seawalls and good agreement has been found. The model requires more verification tests for irregular waves before its application as a generic design tool.     

基于单景高分辨率光学卫星遥感影像的浅海水下地形提取

本文基于海浪波折射现象和浅水波理论,提出了一种基于单景高分辨率光学遥感影像的浅海地形提取方法。首先,基于浅水波理论推导出适用于浅海区域的水深与海浪波长、频率的定量关系,针对近岸光学遥感图像复杂的海浪特征,讨论了两种海浪波长提取方法,即FFT方法和剖面线法。然后提出了基于长距离波长波动分析的海浪频率计算方法,解决了单景遥感影像的波浪频率计算难题。最后,利用单景QuickBird高分辨率光学遥感影像,以海南岛三亚湾为研究区域进行了应用实验,结果表明,对12m以浅的浅海区域,在不需要任何辅助参数的情况下,反演获得了浅海地形(DEM),经与1:25000比例尺海图的水深对比验证,地形趋势吻合良好,反演水深的均方根误差为1.07m,相对水深误差为16.2%,表明该方法适合于浅海水下地形的提取,且具有无需实测水深数据和环境参数的支持的优点。     

Data assimilation with a barotropically unstable shallow water system using representer algorithms

The cycling representer algorithm of Xu and Daley (2000) is a weak constraint four-dimensional variational data assimilation algorithm. It was successfully applied to a one-dimensional transport problem and was able to successfully extract the signal from noisy and sparse observations. The algorithm, however, has not previously been applied to a multivariate, multidimensional system with dynamic instability. The algorithm is also very computationally demanding and awaits considerable enhancement in computer power before being practical for operational forecast models. We have two objectives in this paper. The first is to apply the cycling representer algorithm to a two-dimensional, multivariate barotropically unstable linear shallow water system. The second objective is to formulate and test an accelerated representer algorithm that is much more computationally tractable than the cycling representer algorithm itself. A linear shallow water system with a barotropically unstable basic state was used as a test bed to conduct data assimilation experiments. The evolution of a 'neutral' eastward-propagating singular vector was selected as the 'truth', against which all data assimilation experiments were to be evaluated. The results indicated that the cycling representer algorithm was capable of providing satisfying state estimates for a multivariate, multidimensional system. The results from the accelerated representer algorithm were very encouraging because it is sufficiently computationally tractable to be used on present day multi-processor machines for operational applications.     

Efficient computation of surf zone waves using the nonlinear shallow water equations with non-hydrostatic pressure

A numerical method for non-hydrostatic, free-surface, irrotational flow governed by the nonlinear shallow water equations including the effects of vertical acceleration is presented at the aim of studying surf zone phenomena. A vertical boundary-fitted grid is used with the water depth divided into a number of layers. A compact finite difference scheme is employed for accurate computation of frequency dispersion requiring a limited vertical resolution and hence, capable of predicting the onset of wave breaking. A novel wet–dry algorithm is applied for a proper handling of moving shoreline. Mass and momentum are strictly conserved at discrete level while the method only dissipates energy in the case of wave breaking. The numerical results are verified with a number of tests and show that the proposed model using two layers without ad-hoc assumptions enables to resolve propagating nonlinear shoaling, breaking waves and wave run-up within the surf and swash zones in an efficient manner.     

Analysis of body supercavitation in shallow water

Motion of a cavitating body in shallow waters undergoes a blockage effect. There are influences of the rigid boundary (the sea bottom) and the free boundary (sea surface) in shallow waters. As shown by computation carried out with the ideal fluid theory, the combination of these influences leads to an increase of cavitation number for a cavity of a fixed length and to 3D deformations of the cavity cross-sections, with a swelling of the down cavity part.     

Tomographic mapping of sediments in shallow water

Sediment compressional wave speeds were estimated using broad-band data in range-dependent environments. The environment was assumed as mildly range dependent and was modeled using adiabatic theory. The inversion scheme was based on group speed-dispersion behavior. A genetic algorithm (GA) combined with a neighborhood approach was used for the search. The top layer of sediment was mapped in the shelf region using acoustic data from explosive sources collected on a vertical line array.     

GEK measurements in a shallow water region

Evidence for the applicability of GEK (Geomagnetic Electrokinetography) measurements to shallow water regions is provided from observations in the shelf region of the East China Sea. The reason for the effectiveness of GEK measurements in this case is investigated theoretically, and it is shown to be attributable to the existence of a thick conductive sedimentary layer. In addition, it is shown that low conductive basement rock can be regarded as a good conductor for GEK measurements if the current width is broad enough and if the ratio of current width to water depth is larger than the resistivity ratio of basement rock to sea water. This implies that barotropic tidal currents can be measured with GEK in any ocean on the earth if they have significant magnitudes.     

Animal-sediment relations in shallow water benthic communities

Ninety benthic samples were used to study animal-sediment relations in Tomales Bay, California. While most of the benthic species studied were found more often in a particular type of substrate, individuals of such species were occasionally found in other substrates. There is a striking tendency for species occurring outside of their characteristic environment to be associated with the most diverse assemblages of the foreign substrate. Species low in the order of succession are those species that are found more frequently on other substrates. These phenomena are explained in terms of environmental stability. The concepts involved suggest a means of predicting the sequence of faunal changes following the alteration of the substrate.     

Extraction of morphometric bedform characteristics from profiling sonar datasets recorded in shallow coastal waters of China

The morphological characteristics of small-scale bedforms were measured by means of an acoustic profiling sonar on the Dafeng tidal flat,Jiangsu,in 2009,and in the Jiulong Estuary,Xiamen,in 2010,respectively.The "multi-threshold value" method was utilized to reveal the morphological undulations along which bedforms were present.Analyses of the datasets obtained show that:(1) sand ripples can have irregular shapes,and(2) changes in bedform morphology are small within a single tidal cycle but may be significant over several tidal cycles.Fractal and variogram analyses of the seabed roughness revealed the existence of a significant relationship between current speed and the fractal dimension of the seabed roughness.As current speed increases,seabed roughness increases with a trend of smaller-scale bottom structures being replaced by larger-scale structures.Furthermore,the surface of the larger-scale bottom structures can either become smooth due to the absence of smaller-scale features or become rougher due to the presence of superimposed smaller-scale structures.