Pseudo‐dynamic testing with substructuring at the ELSA Laboratory

The paper describes the distinctive features of the pseudo‐dynamic test method as implemented at the ELSA reaction‐wall facility. Both hardware and software aspects are considered. Particular attention is devoted to the digital control system and to a coupled numerical–experimental substructuring technique allowing realistic earthquake testing of very large structures. Mathematical and implementation details corresponding to this testing technique are given for both synchronous and asynchronous input motion. Selected test results illustrate the advantages of the presented features. Copyright © 2000 John Wiley & Sons, Ltd.     

引入PLSA模型的光学遥感图像舰船检测

提出一种基于概率潜在语义分析(Probabilistic Latent Semantic Analysis,PLSA)的检测算法,首先通过PLSA将目标表述为潜在成分的概率组合,然后利用统计模式识别方法对获取的潜在成分概率进行判别,从而完成最终的检测。其中,生成的潜在成分反映了目标与特征之间相互出现的频率关系,并以潜在成分在目标中概率差异的形式对上述不对应现象给出了直观描述。实验结果表明,所提出的算法对多种复杂情况下的光学图像舰船检测具有很好的准确性和鲁棒性。     

高分辨率航空影像建筑物垂直边缘线提取技术研究

高分辨率航空影像建筑物垂直边缘线的提取在获取建筑物高度信息和检测建筑物过程中可以发挥重要的作用。分析建筑物垂直边缘线在遥感影像上的特性,在遥感影像边缘检测和细化基础上,利用带约束条件的Hough变换实现建筑物垂直边缘线的提取,通过实验验证该方法的可行性。     

一种模拟非均匀介质中弹性波传播新的k-space方法

传统的伪谱(PS)方法,采用傅里叶变换(FT)计算空间导数具有很高的精度,每个波长仅需要两个采样点,而时间导数采用有限差分(FD)近似因而精度较低.当采用大时间步长时,由于时空精度不平衡,PS法存在不稳定性问题.原始的k-space方法可以有效地克服这些问题但是却无法适用于非均匀介质.为了提高原始k-space方法模拟非均匀介质波动方程的精度,我们提出了一种新的k-space算子族.它是用非均匀介质的变速度代替原k-space算子中的常数补偿速度构造得到,引入低秩近似可以高效求解.我们将构造的新的k-space算子应用于耦合的二阶位移波动方程,而不是交错网格一阶速度应力波动方程,使模拟弹性波的计算存储量减少.我们从数学上证明了基于二阶波动方程的k-space方法与基于一阶波动方程的k-space方法是等价的.数值模拟实验表明,与传统的PS、交错网格PS和原始的k-space方法相比,我们的新方法可以在时间和空间步长较大的均匀和非均匀介质中,为弹性波的传播提供更精确的数值解.在保持稳定性和精度的同时,采用较大的时空采样间隔,可以大大降低数值模拟的计算成本.     

基于两步弥散算子的共反射面元射线束成像方法研究

基于克希霍夫统一成像理论,对共反射面元射线束成像方法进行了深入分析.通过这种分析可以得到一种新的共反射面元射线束类成像实现方式:基于两步弥散算子的共反射面元射线束类成像方法.当仅实施第一步弥散时,即可获得信噪比大幅提高同时更为规则的叠前道集.二维、三维的理论和实际数据算例证实了上述观点.     

基于逆算子估计的AVO反演方法研究

传统反演算法以优化算法为主,而基于逆算子估计的AVO反演算法则利用了直接求逆的思路.算法的关键在于寻找存在逆函数的子域,进而可以在子域内直接求逆,这种解决反问题的思路不同于一般的优化类算法所采用的直接搜索解的方式,具有更高的效率.AVO反演利用了振幅随着偏移距的变化特征,反演的精度受到地震资料质量的影响,通过加入L1范数约束以及合理的初始模型有助于提高反演的稳定性以及准确度.模型测算和实际应用表明,基于逆算子估计的AVO反演方法具有较高的精确程度和可靠性.     

基于截断牛顿法的VTI介质声波多参数全波形反演

不同类别参数间的相互耦合使多参数地震全波形反演的非线性程度显著增加,地震波速度与各向异性参数取值数量级的巨大差异也会使反演问题的性态变差.合理使用Hessian逆算子可以减弱这两类问题对反演的影响,提高多参数反演的精度,而截断牛顿法是一种可以比较准确地估计Hessian逆算子的优化方法.本文采用截断牛顿法在时间域进行了VTI介质的声波双参数同时反演的研究.不同模型的反演试验表明,在VTI介质声波双参数同时反演中,截断牛顿法比有限内存BFGS(Limited-memory Broyden-Fletcher-Goldfarb-Shanno,L-BFGS)法能更准确地估计Hessian逆算子,进而较好地平衡两类不同参数的同时更新,得到了比较精确的反演结果.     

Generalized diffraction‐stack migration and filtering of coherent noise

We reformulate the equation of reverse‐time migration so that it can be interpreted as summing data along a series of hyperbola‐like curves, each one representing a different type of event such as a reflection or multiple. This is a generalization of the familiar diffraction‐stack migration algorithm where the migration image at a point is computed by the sum of trace amplitudes along an appropriate hyperbola‐like curve. Instead of summing along the curve associated with the primary reflection, the sum is over all scattering events and so this method is named generalized diffraction‐stack migration. This formulation leads to filters that can be applied to the generalized diffraction‐stack migration operator to mitigate coherent migration artefacts due to, e.g., crosstalk and aliasing. Results with both synthetic and field data show that generalized diffraction‐stack migration images have fewer artefacts than those computed by the standard reverse‐time migration algorithm. The main drawback is that generalized diffraction‐stack migration is much more memory intensive and I/O limited than the standard reverse‐time migration method.     

Uncertainty assessment of ephemeral gully identification,characteristics and topographic threshold when using aerial photographs in agricultural settings

Manual digitizing on aerial photographs is still commonly used for characterizing gully erosion over large areas. Even when automated detection procedures are implemented, manual digitizing is frequently being resorted to in order to constitute reference datasets used for training and validation. In both cases, manual digitizing entails some subjective decisions on behalf of the operator, which introduces uncertainty into the resulting datasets. To assess the magnitude of this uncertainty, 11 experienced operators were asked to digitize and classify ephemeral gullies (EGs) on cropland following a standardized methodology. The resulting 11 datasets were compared in terms of number, type and location of EGs. Furthermore, for EGs located on a well‐defined runoff flow concentration axis, the slope versus contributing area topographic thresholds required for initiating gully channels were assessed using four thresholding methods, and compared across the 11 datasets. The operators identified 259 different EGs. However, the number (52–139) and sum total length (8.9–23.7 km) of EGs varied widely across operators. Only 34% of the EGs were digitized by more than half of the operators, and 7% were identified by all. Identification of EGs located on a well‐defined flow concentration axis proved least subjective. The longer the EG and the more fields the EG crossed, the larger the number of operators that were able to identify it. EGs were also most easily identified when located in sugar beet fields as compared to other crops. EG classification and topographic threshold lines were also found to be strongly operator‐dependent. Quantile regression appeared to be one of the most robust thresholding methods. Operator subjectivity when digitizing EGs on orthophotographs introduces uncertainty that should be taken into account in future remote sensing‐based studies of EG erosion whenever they rely, in part or in full, on manual photograph interpretation. Copyright © 2014 John Wiley & Sons, Ltd.     

Representations of inverse covariances by differential operators

In the cost function of three- or four-dimensional variational data assimilation, each term is weighted by the inverse of its associated error covariance matrix and the background error covariance matrix is usually much larger than the other covariance matrices. Although the background error covariances are traditionally normalized and parameterized by simple smooth homogeneous correlation functions, the covariance matrices constructed from these correlation functions are often too large to be inverted or even manipulated. It is thus desirable to find direct representations of the inverses of background error correlations. This problem is studied in this paper. In particular, it is shown that the background term can be written into f dx|Dv(x)|^2, that is, a squared L2 norm of a vector differential operator D, called the D-operator, applied to the field of analysis increment v(x). For autoregressive correlation functions, the D-operators are of finite orders. For Gaussian correlation functions, the D-operators are of infinite order. For practical applications, the Gaussian D-operators must be truncated to finite orders. The truncation errors are found to be small even when the Gaussian D-operators are truncated to low orders. With a truncated D-operator, the background term can be easily constructed with neither inversion nor direct calculation of the covariance matrix. D-operators are also derived for non-Gaussian correlations and transformed into non-isotropic forms.