一种四阶P-Laplace图像盲复原方法

针对TV图像复原方法中产生的阶梯效应问题,结合传统的差值范数图像复原方法,提出一种四阶P-Laplace图像盲复原算法。该算法将四阶偏微分方程引入P-Laplace图像复原方法中,对图像的边缘和内部平滑区域做了分别处理以缓解TV复原方法中带来的阶梯效应,同时引入参数λ调节复原过程中的保真项,并对偏微分方程做单调灰度变化,进一步改善了图像复原效果,缓解了阶梯效应。实验结果证明,该算法对模糊图像有很好的复原效果,也适用于水下图像和雾天图像的图像复原。     

水下图像基于 GAN 去模糊的增强技术

由于水下环境具有不稳定性,水下图像可能会出现偏色、对比度低以及运动模糊等退化现象。针对这些问题,本文提出了适用于水下图像的增强算法,其实现需要依次经过颜色恢复和去模糊这2个阶段。在第一阶段中,本文增强算法先利用高斯滤波和均值漂移对图像进行锐化;然后,通过对比图像各颜色通道的均值得到补偿值对图像颜色进行校正;最后通过线性拉伸来调整图像的对比度。在第二阶段中,采用带有残差思想的生成对抗网络(GenerativeAdversarialNetworks,GAN),利用9个连续的残差网络能够很好地提取图像中的特征,可起到消除模糊和增强图像特征的作用。利用本文算法处理水下图像时,发现本文方法不仅能去除图像模糊,而且能消除图像的色偏现象且不携带红色伪影。同时,通过对比水下图像质量度量(UnderwaterImageQualityMeasures,UIQM)和水下彩色图像质量评估(UnderwaterColorImageQualityEvaluation,UCIQE)这2项指标发现,本文算法有较好的图像处理效果。     

多种群并行进化神经网络的研究及应用

提出一种新的多种群并行遗传算法 (NMPGA) ,并将其作为多层前馈神经网络(MFNNs)的学习算法 ,从而形成一类新的 MFNN模型——多种群并行进化神经网络(MPENNs)。首先 ,对一给定的网络结构 ,随机产生一初始权重的集合 ,这个集合实际上对应着一组具有相同结构但不同权重的神经网络。然后 ,采用 NMPGA对 MFNNs的权重进行进化。最后 ,性能最好的网络被选作目标问题的解。在 NMPGA算法中 ,作者采用浮点数编码来克服传统二进制编码的精度不足问题 ,并设计了专门的杂交算子和变异算子来增强算法性能。实验结果表明 ,MPENNs能成功解决异或问题、三元奇偶问题及成品烟的感官质量评价问题。     

基于神经网络的规则提取及其渔业应用研究

为了解决远洋渔业中过度依赖经验而产生的盲目捕捞问题,结合海洋环境数据和历史产量数据对渔场进行有效分析,提出了一种基于径向基函数神经网络(Radial basis function neural network,RBFNN)的栖息地指数(HSI)预测方法,并将其应用于印度洋海域大眼金枪鱼(Thunnus obesus)栖息地指数的预测。在RBFNN训练过程中使用模糊C均值(Fuzzy c-means,FCM)聚类算法,在基于神经网络的规则提取过程中首次采用了和声搜索(Harmony search,HS)算法。实验研究表明,利用FCM改进后的RBFNN,均方误差(Mean square error,MSE)达到0.021 6。和声搜索由于算法简单,易于实现,能够应用于训练后的FCM-RBFNN提取分类规则,提取出的规则能够反映该渔业现状。     

基于DPHP技术测量热导率的简化算法

双针热脉冲(DPHP)技术在海底沉积物热导率的原位测量中有很好的应用前景,但其算法复杂,而且存在累积误差。通过对DPHP技术理论模型的分析,提出了一个简化算法,探讨了算法的理论依据和适用条件。对实测数据进行运算的结果表明,该算法能够有效简化热导率的计算,并且误差在合理范围内。     

基于模糊C均值聚类-支持向量机的海底沉积物分类识别

在总结了目前海底底质分类研究的基础之上,率先提出利用计算机数值模拟技术对海底底质进行分类识别研究。相较于目前海底底质分类研究中所使用的水槽实验法,提出采用计算机数值正演技术模拟实际地震勘探中数据采集过程。在分类识别算法上,分别采用支持向量机(SVM)和模糊C均值聚类(FCM)算法对采集的数据进行分类,为使支持向量机分类识别率达到最大,引入差分进化算法对支持向量机中关键参数进行最优化搜索,并研究了向原始地震记录中加入10%,30%,50%的高斯白噪音时算法的稳定性。在分析了这两种算法分类识别的正确率及其各自的优缺点后,提出了海底底质分类识别的两步法,即(1)先利用模糊C均值聚类进行一粗糙的预测分类,在每一类中挑选聚类性较好的数据作为支持向量机的训练样本;(2)将上一步中筛选的样本作为支持向量机的训练样本,并用差分进化算法优化支持向量机分类参数,再利用训练好的支持向量机对其余数据做预测分类。鉴于计算机数值模拟的可重复性、高效快速性及本文提出的模糊C均值聚类-支持向量机方法的鲁棒性,为便于开展进一步研究,归纳总结了一套行之有效的采用计算机数值模拟技术开展海底底质分类识别研究的一般化流程。     

一种基于TIN模型的LOD简化算法

为了实现不规则三角网（TIN）地形模型的动态模型的动态细节简化模型,本文提出了一种有效的方法,即在TIN的凸壳构网算法和累进格网算法的基础上,设计一种混合型数据结构,提出边折叠的限制条件,解决了边折叠时易产生的边自交现象,实现了TIN模型的LOD快速简化。     

Bancroft算法应用于GPS/BDS/GLONASS三系统伪距单点定位

为了提高多系统组合单点定位的运算效率,将Bancroft算法推广应用于GPS/BDS/GLONASS三系统伪距单点定位,并推导了该方法的详细计算流程,比较了Bancroft算法、最小二乘算法和组合两种算法的精度、迭代次数以及运算时间。利用实测数据进行实验,结果表明,3种算法精度相当,Bancroft算法比最小二乘算法迭代次数少,运算效率提高了38%,利用Bancroft为最小二乘算法提供初值,收敛速度加快,运算效率提高了36%。最终得出结论,Bancroft算法应用于GPS/BDS/GLONASS三系统伪距单点定位满足精度要求,且Bancroft算法对于减少迭代次数,提高运算效率有重要意义。     

GPS系统下RAIM算法可用性及结果分析

介绍了接收机自主完整性监测(RAIM)算法的故障检测及其可用性原理,结合航空用户对完好性需求,通过仿真计算,得到了GPS系统下不同截止高度角条件下、不同飞行阶段的RAIM算法的可用性结果,并且给出了中国区域内5°×5°格网点的RAIM最大和平均中断时间,分析了截止高度角、可视卫星个数和DOP值等因素对RAIM算法可用性的影响。     

振动系统的逆与优化问题(Part-Ⅱ)部分边界值再构内部场(I)理论(英文)

研究由一振动体的部分边界测量数据再构振动体内部场的逆问题。提出基于偏微分方程控制方法的一种算法 ,并证明原用于处理散射问题的 Null- field method可用来求解优化问题的随伴问题。     

振荡剪切流作用下钢悬链线立管非线性动力分析

钢悬链线立管(SCR)的非线性动力响应会在复杂的海洋环境下发生改变。以钢悬链线立管为研究对象,采用细长柔性杆理论,考虑波浪与振荡剪切流联合作用,建立钢悬链线立管运动方程并进行有限元离散,在时域中求解,编写相应计算程序,通过算例研究钢悬链线立管在波浪与振荡剪切流联合作用下的动力响应变化规律。将振荡剪切流与一般剪切流进行对比,得到突发海况下振荡剪切流对钢悬链线立管动力响应的影响规律,为复杂海洋环境下的工程设计提供依据。     

Time series prediction of mooring line top tension by the NARX and Volterra model

To minimize the computation burdens of long-term analysis, the mooring line top tensions under various short-term sea states were predicted using the pre-calculated nonlinear time domain analysis results of some selected sea states. A nonlinear autoregressive with an exogenous input (NARX) technique was introduced with a finite nonlinear memory length to predict the top-tension of a mooring line of the box-type floating production storage and offloading (FPSO). The NARX was designed in such a way that it takes the floater’s motions as its input and produces the mooring line top tension as an output. After training the NARX using the pre-calculated time series of the motions and tensions of some selected sea states, the tension time histories of different sea sates were predicted and compared with the direct time domain analysis results. In addition, to explore the nonlinearity of the system and its contribution to the response, the transfer functions of different orders were extracted after expanding the NARX equation using a Taylor series expansion. The nonlinear contributions coming from both the 3^rd and 5^th order were significant.     

黄、东海环流的数值研究Ⅱ潮及潮致环流数值模拟

该文中数值模拟了黄、东海陆架区的四大天文潮及其潮致拉格朗日环流。运用“流速分解法”将线性三维分潮波问题化为一个分潮潮位及深度平均流的二维问题和一个流速垂向剖面问题求解。其二维问题的底摩擦系数由三维模式导出 ,所导得的线性底摩擦力公式更符合物理意义 ,且避免一般二维模式中底摩擦力的经验假定。结果表明 ,三维模式完全可以通过准二维方法求解 ,以避免一般三维问题的繁重计算     

水下模糊图象维纳滤波和卡尔曼滤波恢复算法研究

该文将海水中图象传输特性的研究与用此传输特性进行水下模糊图象恢复结合起来 ,得到一套测量水体传递函数及进行水下图象处理的方法。采用维纳 (Wiener)滤波和卡尔曼 (Kalman)滤波算法 ,对模糊图象进了恢复。结果表明 ,利用计算机采用一定的算法对散射性质已知的水体造成的图象模糊可以进行有效的复原 ,且良好效果。     

Comparative Study of Different SPH Schemes on Simulating Violent Water Wave Impact Flows

Free surface flows are of significant interest in Computational Fluid Dynamics（CFD）. However, violent water wave impact simulation especially when free surface breaks or impacts on solid wall can be a big challenge for many CFD techniques. Smoothed Particle Hydrodynamics（SPH） has been reported as a robust and reliable method for simulating violent free surface flows. Weakly compressible SPH（WCSPH） uses an equation of state with a large sound speed, and the results of the WCSPH can induce a noisy pressure field and spurious oscillation of pressure in time history for wave impact problem simulation. As a remedy, the truly incompressible SPH（ISPH） technique was introduced, which uses a pressure Poisson equation to calculate the pressure. Although the pressure distribution in the whole field obtained by ISPH is smooth, the stability of the techniques is still an open discussion. In this paper, a new free surface identification scheme and solid boundary handling method are introduced to improve the accuracy of ISPH. This modified ISPH is used to study dam breaking flow and violent tank sloshing flows. On the comparative study of WCSPH and ISPH, the accuracy and efficiency are assessed and the results are compared with the experimental data.     

Numerical Prediction of Vortex Induced Vibrations on Top Tensioned Riser in Consideration of Internal Flow

In consideration of the effect of the internal flowing fluid and the external marine environmental condition on the vortex-induced vibration （VIV） of top tensioned riser （Till）, the differential equation is derived based on work-energy principles and the riser near wake dynamics is modeled by Facchinetti＇ s wake oscillator model. Then Galerkin＇ s finite element approximation is implemented to derive the nonlinear matrix equation of the coupled equations and file corresponding numerical programs are compiled which solve the coupled equations directly in the time domain. The comparison of the predicted results with the recent experimental results and the prediction of SHEAR7 is performed. The results show the validity of the proposed method on the prediction of VIV of deep water risers. The effect of internal flow on the dynamic characteristics and dynmnic response of the riser is analyzed and several valuable conelusions are drawn.     

Nonlinear Effect of Wave Propagation in Shallow Water

—In this paper,a nonlinear model is presented to describe wave transformation in shallow wat-er with the zero-vorticity equation of wave-number vector and energy conservation equation.Thenonlinear effect due to an empirical dispersion relation(by Hedges)is compared with that of Dalrymple＇sdispersion relation.The model is tested against the laboratory measurements for the case of a submergedelliptical shoal on a slope beach,where both refraction and diffraction are significant.The computation re-sults,compared with those obtained through linear dispersion relation.show that the nonlinear effect ofwave transformation in shallow water is important.And the empirical dispersion relation is suitable for re-searching the nonlinearity of wave in shallow water.     

任意曲线边界条件下缓变水深水域波浪传播的数值模拟

缓坡方程被广泛地应用于描述波浪的传播变形计算,目前一般采用矩形网格求解.将计算域剖分为任意四边形网格,以格林公式为基础,在变量沿单元边界线性变化的假定下,对双曲型的波能守恒方程、波数矢无旋性方程进行离散,同时通过等参单元变换推求节点偏导数值以离散椭圆型光程函数方程,从而建立了任意曲线边界条件下缓变水深水域波浪传播的数值模拟模型.将模型应用于平行直线型等深线地形,并将计算域剖分为不规则四边形网格,对不同入射角、底坡、波高等多种组合情况比较了数值解与解析解,结果表明两者一致.应用于复杂边界的实例,数值模拟结果与物模实验值基本吻合.     

河口往复流中潮流不对称与推移质输沙的关系

潮流不对称现象是近岸潮波运动的基本特征之一,为了研究潮流不对称对推移质泥沙长期净输运的作用机制,首先在往复流情况下对Bagnold推移质输沙率公式(1966)由于不包含起动流速所可能导致的误差进行理论分析,进而由统计学中的"偏度"的概念出发,推导了河口潮流不对称与推移质输沙之间的定量关系。研究认为,虽然Bagnold推移质输沙率公式不包含泥沙起动流速,但是只要泥沙起动流速和最大流速的比值在一定范围内,则在理论上公式的计算误差是可以接受的。由偏度结合Bagnold公式导出了潮流不对称与推移质输沙之间的关系,表明推移质泥沙的长期净输运不仅与余流有关,而且与不同分潮组合之间、余流与分潮之间的相互作用紧密相关。该关系还给出了一个由潮流调和常数估算河口推移质输沙的简便方法。经对比,从潮流不对称出发估算推移质输沙与直接采用推移质输沙率公式结果一致。     

A novel decomposition of the quadratic transfer function (QTF) for the time-domain simulation of non-linear wave forces on floating bodies

In the present study, a novel method is proposed for the separation of the second-order sum- and difference-frequency wave forces—that is, quadratic transfer functions (QTFs)—on a floating body into three components due to wave–wave, wave–motion, and motion–motion action. By applying the new QTF components, the second-order wave forces on a floating body can be strictly computed in the time domain. In this work, the boundary value problems (BVPs) corresponding to the three kinds of QTF components were derived, and non-homogeneous boundary conditions on the free surface and the body surface were obtained. The second-order diffraction potentials were determined using the boundary integral equation method. In the solution procedure, the highly oscillatory and slowly converging integral on the free surface was evaluated in an accurate and effective manner. Furthermore, the application of the QTF components in the time domain was demonstrated. The second-order exciting forces in the time domain were divided into three parts. Each part of these forces was computed via a two-term Volterra series model based on the incident waves, the first-order motion response, and the QTF components. This method was applied to several numerical examples. The results demonstrated that this decomposition yields satisfactory results.