小波多尺度影像边缘聚焦算法

讨论了基于光滑函数的二维小波极大模边缘算子和小波过零点边缘算子及其特性。提出了多尺度二进小波变换过零点边缘聚焦算法。通过对不同尺度下小波变换的遥感影像进行复合、边缘提取和聚焦后结果进行比较，可以发现该算法适应影像空间尺度不确定性目标边缘提取。既可保证大尺度下的轮廓信息不失真，又能在边缘定位上保持很高的精度。     

Automatic spikes detection in seismogram

Data processing for seismic network is very complex and fussy, because a lot of data is recorded in seismic network every day, which make it impossible to process these data all by manual work. Therefore, seismic data should be processed automatically to produce a initial results about events detection and location. Afterwards, these results are reviewed and modified by analyst. In automatic processing data quality checking is important. There are three main problem data thatexist in real seismic records, which include: spike, repeated data and dropouts. Spike is defined as isolated large amplitude point; the other two problem datahave the same features that amplitude of sample points are uniform in a interval. In data quality checking, the first step is to detect and statistic problem data in a data segment, if percent of problem data exceed a threshold, then the whole data segment is masked and not be processed in the later process.     

截断误差最小化的优化空间有限差分算子

气候和天气数值预测本质上都需建立在离散空间的基础上，并要求有高精度的空间离散化计算方案。然而，解析性微分算子和离散性差分算子的差异却是造成数值计算的原始误差源。作根据约束性代价函数原理提出了能够使得截断误差最小化的中心差分优化算子。从最优化角度看，该算子本质上是以一种简单但却相当创新的方式尝试应用变分连续同化（VCA）技术来最小化大气模式中的空间截断误差。此外，该优化差分算子的设计不依赖于网格结构，即可应用在绝大多数Arakawa网格上，诸如非跳点网格（Arakawa-A)以及常规跳点网格（Arakawa-B,Arakawa-C,Arakawa-D)。但为了对优化差分算子精确性进行严格测定，该算子的基准试验结果是通过在截断误差最易被激发出来的非跳点网格上进行严格理论和实际数值模拟测试得到的。两的结果都表明该优化差分算子截断误差造成的数字噪音能被减小到最小量。     

受迫Liénard方程周期解的存在性

对于受迫Liénard方程,利用Sobolev空间范数,给出了周期解的估计,进而利用变分原理、Schauder不动点定理,证明了周期解的存在性。     

粘弹性圆形巷道支护结构上的荷载及其围岩应力的解析解

叙述了采用积分算子解法时弹性解与粘弹性解之间的对应关系;具体求算出H-K体的蠕变核与松弛核;借助带有弹性支护的粘弹性围岩应力解析解,计算分析了支护结构上荷载以及围岩应力的分布形式与变化规律,对围岩的应力调整有了更深一步的认识,以便更好地为衬砌设计服务。     

地震偏移的最优可分近似算法实现

针对地震偏移算法中单程波算子的特征函数近似，采用最优可分表示法将该特征函数展开为空间变量(g)和水平波数(k)的可分表达式，以此可分近似表达式为基础，运用正反Fourier变换重新构造单程波算子. 为了克服特征函数最优可分近似计算在奇点及其领域产生的数值振荡，引入等价黏性技巧以增强算法的数值计算稳定性，并采用分频最优可分及对空间变量(g)线性插值的方法，不仅提高了计算精度，也节约计算机时. 文中具体研究了单程波算子的脉冲响应和二维叠前深度偏移. 结果表明，在不甚大步长情况下，本文构造的算子具有适应横向强变速的能力，运用Marmousi模型验证了本文方法适合复杂构造的成像.     

Joint estimation of spatial deformation and blurring in environmental data

Joint estimation of deformation and blurring parameters from spatial observations is considered. The generalized random field approach to the problem introduced in this paper provides a suitable framework for a technical treatment of these effects in relation to singularity properties of the fields. A mixture of Kullback-Leibler divergence loss functions is formulated in a Bayesian context. Simulations are developed to illustrate the performance of the approach proposed.     

基于扩展集合操作的频繁项集挖掘算法研究

对传统集合操作进行了扩展，提出了基于扩展集合操作的最大频繁项集生成算法FIS-ES，并从理论上对算法的复杂度进行了详细的分析。实验表明，在最小支持度较小的情况下，FIS-ES算法比Apriori算法具有更快的挖掘速度、更少的空间占用等优点，与Apriori有很好的互补性。     

3D wavefield extrapolation with optimal separable approximation

An accurate and wide-angle one-way propagator for wavefield extrapolation is an important topic for research on wave-equation prestack depth migration in the presence of large and rapid velocity variations. Based on the optimal separable approximation presented in this paper, the mixed domain algorithm with forward and inverse Fourier transforms is used to construct the 3D one-way wavefield extrapolation operator. This operator separates variables in the wavenumber and spatial domains. The phase shift operation is implemented in the wavenumber domain while the time delay for lateral velocity variation is corrected in the spatial domain. The impulse responses of the one-way wave operator show that the numeric computation is consistent with the theoretical value for each velocity, revealing that the operator constructed with the optimal separable approximation can be applied to lateral velocity variations for the case of small steps. Imaging results of the SEG/EAGE model and field data indicate that the new method can be used to image complex structure.