Landsat7 ETM＋影像的融合和自动分类研究

利用SFIM、MLT、HPF和修改的Brovey(MB)等遥感影像融合算法对Landsat 7 ETM 影像进行融合和自动分类研究，并就融合影像的光谱保真度、高频空间信息融人度和分类精度对这些方法进行评价。结果表明SFIM变换几乎完全保持了原始影像的光谱特点，并具有最高的平均分类精度；MB变换具有最高的高频空间信息融人度；MLT变换也具有较高的分类精度；只有HPF变换的各项指标都不突出。所有4种融合影像的分类精度都较原始影像的分类精度有明显的提高。这表明，源于同一传感器系统的不同分辨率影像的融合可以避免异源传感器融合影像所常见的各种参数、时相和配准误差，所以能够明显地提高影像的自动分类精度。     

A recursive decomposition algorithm for network seismic reliability evaluation

A new probabilistic analytical approach to evaluate seismic system reliability of large lifeline systems is presented in this paper. The algorithm takes the shortest path from the source to the terminal of a node weight or edge weight network as decomposition policy, using the Boolean laws of set operation and probabilistic operation principal, a recursive decomposition process then could be constructed. For a general weight network, the modified Torrieri method (NTR/T method) is introduced to combine with the suggested algorithm. Therefore, the recursive decomposition algorithm may be applied to evaluate the seismic reliability of general lifeline systems. A series of case studies, including a practical district electric power network system and a large urban water supply system, show that the suggested algorithm supplies a useful probabilistic analysis means for the seismic reliability evaluation of large lifeline systems. Copyright © 2002 John Wiley & Sons, Ltd.     

应用CP网络进行岩性识别

为通过测井解决岩性识别问题，引入了具有分类准确、算法简练等优点的CP（Counter-Propagation)网络。在详细介绍CP网络的网络模型和算法的基础上，结合某油田的实际测井资料，进行了CP网络识别研究。应用结果表明：CP网络训练周期短、识别准确率高、不存在收敛问题。通过试验研究得出结论：CP网络完全可以用于解决岩性识别等问题，具有广阔的应用前景。     

遗传算法在边坡数值计算中的应用

改进了进化方向的遗传算法与有限元数值法，结合并研制了相应的软件。应用该软件对多类型岩土边坡进行弹性模量、内聚力、内摩擦角等参数反演分析，显示误差很小，收敛速度也很快，这说明改进进化方向遗传算法这种新型的优化算法在多类型岩土参数优化估计中具有独特的优势。     

Application of genetic BP network to discriminating earthquakes and explosions

Introduction Artificial Neural Network (ANN) is an important branch of artificial intelligence. It is proposed on the foundation of the study on modern neural science, is a man-made network that can implement some functions based on the mans comprehensive understanding for cerebral neural network (HAN, WANG, 1997). ANN is a mathematical model of simplified human brain neural network and is used to simulate the structures and functions of human brain neural network. ANN is a complex netw…     

Hydration and refractive indices of Holocene tephra glass in Hokkaido, Northern Japan

Refractive indices of incompletely hydrated tephra glasses vary widely. Thus, glass refractive index is not a practical indicator for identifying Holocene tephras. Hydration near the surface of tephra glass shards can be removed by either hydrofluoric acid treatment or annealing for 12 h at 400°C. The annealing procedure is a particularly reliable and simple way to dehydrate tephra glasses. Standard deviations (s) of refractive indices for glasses after 12 h annealing are small (s=0.0014–0.0018), in contrast with untreated glasses (s=0.0034–0.0405). The refractive indices of dehydrated tephra glasses are 0.006–0.014 lower than those of the untreated (hydrated) glasses. Using the 400°C 12-h annealing procedure, values for the refractive indices of eight Holocene tephras in Hokkaido were obtained. These refractive indices are useful indicators for identification and correlation of Holocene tephras.     

Vegetation influences on water yields from grassland and shrubland ecosystems in the Chihuahuan Desert

This study examines runoff generated under simulated rainfall on Summerford bajada in the Jornada Basin, New Mexico, USA. Forty‐five simulation experiments were conducted on 1 m² and 2 m² runoff plots on grassland, degraded grassland, shrub and intershrub environments located in grassland and shrubland communities. Average hydrographs generated for each environment show that runoff originates earlier on the vegetated plots than on the unvegetated plots. This early generation of runoff is attributed to soil infiltration rates being overwhelmed by the rapid concentration of water at the base of plants by stemflow. Hydrographs from the degraded grassland and intershrub plots rise continuously throughout the 30 min simulation events indicating that these plots do not achieve equilibrium runoff. This continuously rising form is attributed to the progressive development of raindrop‐induced surface seals. Most grassland and shrub plots level out after the initial early rise indicating equilibrium runoff is achieved. Some shrub plots, however, display a decline in discharge after the early rise. The delayed infiltration of water into macropores beneath shrubs with vegetation in their understories is proposed to explain this declining form. Water yields predicted at the community level indicate that the shrubland sheds 150 per cent more water for a given storm event than the grassland. Copyright © 2002 John Wiley & Sons, Ltd.     

三维地震数据离散光滑插值的共轭梯度法

针对三维地震数据插值,提出采用Laplacian算子进行光滑约束的插值方法,并借鉴Mallet研究的离散光滑插值思路,采用预条件共轭梯度法,直接生成网格节点上的值,从而回避寻求满足插值方程的函数. 为了实现其中Laplacian算子的快速求逆,文中引入Claerbout螺旋坐标系谱因式分解理论. 在螺旋坐标系下,Laplacian算子的表示矩阵具有Toeplitz结构,其快速求逆可由谱法LU分解实现. 基于二维离散光滑插值,文中还给出共轭梯度法与NMO相结合的沿时间切片逐层处理的离散光滑插值流程. 最后,应用该方法对模型数据和实际三维地震数据进行了处理.     

地震静校正全局最优化问题的求解

针对地震静校正存在的非线性和多参数性等问题，本文提出了一种全局优化的剩余静校正方法.在模型更新规则上，把模拟退火方法和均匀优化设计方法结合在一起，使模型的更新更为合理，从而加快了搜寻全局最优解的速度.同时提出了模型的不同分量的温度参数和退火过程的选取方式，把温度参数的选取与地震剖面的能量联系起来，使温度参数的选取具有自适应的特点.该方法克服了常规模拟退火方法所具有的寻优空间不均匀以及退火参数需通过多次试验选取的缺陷.通过实际地震资料的计算证明，本文提出的方法对地震静校正问题合理而有效.     

大型维歇特地震仪的记录特征及其在地震速报中的作用

大型维歇特地震仪自德国引进南京地震台已近70年。本文简要叙述了该仪器的性能；并以南京地震台的大型维歇特地震仪的记录为例，介绍了该仪器在300-2000km范围内的记录特征，回顾了该仪器在地震速报中所起的历史作用。