面向认知表现预测的时-空共变混合深度学习模型

认知表现预测已经成为当前大脑研究的重要课题.功能磁共振成像技术由于同时具有较好的时间和空间分辨率,有潜力为认知表现预测提供数据支持.为了解决基于功能磁共振成像数据对认知表现进行预测时大脑所具有的时-空共变难刻画问题,提出了一种新型基于大脑学习机制的时-空共变混合深度学习模型,即深度稀疏自编码器与循环全连接网络混合模型,以混合神经网络模型的损失函数误差作为认知表现预测能力的评价标准.在人类连接组项目数据集上的实验结果表明,提出的时-空共变混合模型能够有效和稳健地预测认知表现,并提取到与人脑学习、记忆相关的有意义的脑影像特征,从而为认知表现预测提供技术支持.

X射线CT模型及重建算法

该文从系统论的观点出发,将由投影重建图像中的变换看成空间变化的动态系统模型。基于所建立的连续模型,首先对传统算法的优劣点作了分析比较,接着介绍了基于小波的多分辨分析在图像层析重建中的运用及发展;基于离散模型,从信息论的角度将级数展开法归于单目标优化成像算法,同时介绍了多目标优化理论并对其中多目标熵优化理论的最新发展作了着重探讨。

像移对星载TDICCD相机成像品质的影响分析

像移是影响星载时间延迟积分电荷耦合器件(TDICCD)相机成像品质的重要因素。为研究像移对TDICCD像质的影响,首先分析了像移引起图像模糊和几何变形的机理,研究了像移降质的辐射和几何方面评价参量。依据TDICCD工作原理,建立了像移影响下TDICCD相机空变降质模型,并以此为基础,计算并分析了不同形式、不同方向上的像移影响下的像质下降情况。分析结果表明,不同方向、不同形式的像移引起的成像品质下降是不同的。文章中提出的空变降质模型可以为后续的仿真及像移降质图像的恢复等提供参考。

基于井轨迹地震层位化的水平井时深关系研究及应用

针对倾斜地层对水平井时深映射的影响,从理论模型的角度分析并论证了常规借用时深关系所导致的井震矛盾,在此基础上提出了基于井轨迹地震层位化的解决方法。该方法将构造约束下的层位时深转换理念引入到井轨迹的时深映射中,首先对井轨迹进行地震层位化处理,然后通过建立地质构造约束下的高精度空变速度模型,将井轨迹的深度域地震层位进行深时映射,最终得到井轨迹在时间域更加客观真实的投影形态。该方法在渤海NB油田的实际应用中取得了良好的效果,对类似构造特征区的井位设计及随钻过程中井轨迹的实时调整亦具有较强的指导意义。

A Vector-based Cellular Automata Model for Simulating Urban Land Use Change

Cellular Automata(CA) is widely used for the simulation of land use changes. This study applied a vector-based CA model to simulate land use change in order to minimize or eliminate the scale sensitivity in traditional raster-based CA model. The cells of vector-based CA model are presented according to the shapes and attributes of geographic entities, and the transition rules of vector-based CA model are improved by taking spatial variables of the study area into consideration. The vector-based CA model is applied to simulate land use changes in downtown of Qidong City, Jiangsu Province, China and its validation is confirmed by the methods of visual assessment and spatial accuracy. The simulation result of vector-based CA model reveals that nearly 75% of newly increased urban cells are located in the northwest and southwest parts of the study area from 2002 to 2007, which is in consistent with real land use map. In addition, the simulation results of the vector-based and raster-based CA models are compared to real land use data and their spatial accuracies are found to be 84.0% and 81.9%, respectively. In conclusion, results from this study indicate that the vector-based CA model is a practical and applicable method for the simulation of urbanization processes.

Gradient descent approach for minimizing dissimilarity measure in log-polarimagery

Log-polar mapping has been proposed as a very appropriate space-variant imaging model in active vision applications.There is no doubt about the importance of translation estimation in active visual tracking.In this paper an approach is presented,and its performances are evaluated.The approach uses a gradient descent for minimizing a dissimilarity measure.The experimental results reveal that this method is efficient for approaching active image translations.

Spatiotemporal interpolation of precipitation across Xinjiang, China using space-time CoKriging

In various environmental studies, geoscience variables not only have the characteristics of time and space, but also are influenced by other variables. Multivariate spatiotemporal variables can improve the accuracy of spatiotemporal estimation. Taking the monthly mean ground observation data of the period 1960–2013 precipitation in the Xinjiang Uygur Autonomous Region, China, the spatiotemporal distribution from January to December in 2013 was respectively estimated by space-time Kriging and space-time CoKriging. Modeling spatiotemporal direct variograms and a cross variogram was a key step in space-time CoKriging. Taking the monthly mean air relative humidity of the same site at the same time as the covariates, the spatiotemporal direct variograms and the spatiotemporal cross variogram of the monthly mean precipitation for the period 1960–2013 were modeled. The experimental results show that the space-time CoKriging reduces the mean square error by 31.46% compared with the space-time ordinary Kriging. The correlation coefficient between the estimated values and the observed values of the space-time CoKriging is 5.07% higher than the one of the space-time ordinary Kriging. Therefore, a space-time CoKriging interpolation with air humidity as a covariate improves the interpolation accuracy.

Diagnostic model construction and example analysis of habitat degradation in enclosed bay: II. spatiotemporal variations in habitat degradation in Sansha Bay

The process of habitat degradation varies in habitat type and driving force which shows certain spatial and temporal heterogeneity on regional scales. In the present study, a new diagnostic model for enclosed bay habitat degradation was established, with which the spatial and temporal variation patterns of habitat degradation during 1991–2012 in Sansha Bay, Fujian, China was investigated. The results show that anthropogenic disturbance is the major controlling factor for the habitat degradation in large temporal heterogeneity in the bay. On the other hand, the habitat degradation experienced signifi cant spatial variations among six sub-bays. Under the joint action of temporal and spatial heterogeneity, the degradation trend in growing scale shows a more signifi cant correlation with the distribution of local leading industries along shorelines. Therefore, we quantifi ed the main characters of habitat degradation in Sansha Bay, and have understood the relationship between the status of habitats spatio-temporal variation value and the main controlling factor leading to the changes. However, a defi ciency of this research is the lack of or inaccessible to the detailed data, which shall be better solved in the future study for accessing more data from more sources.

Spatial and temporal variation of drought index in a typical steep alpine terrain in Hengduan Mountains

This study describes the spatial and temporal variation of a drought index and makes inferences regarding the environmental factors that influence this variability in the Hengduan Mountains. A drought index is typically used to determine the moisture conditions and the magnitude of water deficiency in a given area. Based on data from 26 meteorological stations over the period 1960-2012, the spatial and temporal variations of the drought index were analyzed using a thin plate smoothing splines method that considered elevation as a covariate. The drought index was estimated based on the potential evapotranspiration(E0) as defined by the Penman Monteith model modified by FAO(1998). The results of the reported analysis showed that the drought index in the Hengduan Mountains has been decreasing since 1960 at a rate of-0.008/a. This represented a progressive shift from the "sub-humid" class, which typified the wider area in the Hengduan Mountains, toward the "humid" class, which appeared in the Hengduan Mountains areas. The drought index was relatively high in the north and low in the south and the variation of the drought index varied with seasons. The drought index showed increasing trends in summer and autumn and it is greater in autumn than in summer, while it showed a decreasing trend in spring and winter. Drought index is inversely proportional to the soil relative humidity and Normalized Difference Vegetation Index(NDVI).

Responses of Tree Species to Climate Warming at Different Spatial Scales

Tree species respond to climate change at multiple scales,such as species physiological response at fine scale and species distribution (quantified by percent area) at broader spatial scale.At a given spatial scale,species physiological response and distribution can be correlated positively or negatively.The consistency of such correlation relationships at different spatial scales determines whether species responses derived from local scales can be extrapo-lated to broader spatial scales.In this study,we used a coupled modeling approach that coupled a plot-level ecosystem process model (LINKAGES) with a spatially explicit landscape model (LANDIS).We investigated species physio-logical responses and distribution responses to climate warming at the local,zonal and landscape scales respectively,and examined how species physiological response and distribution correlated at each corresponding scale and whether the correlations were consistent among these scales.The results indicate that for zonal and warming-sensitive species,the correlations between species physiological response and distribution are consistent at these spatial scales,and therefore the research results of vegetation response to climate warming at the local scale can be extrapolated to the zonal and landscape scales.By contrast,for zonal and warming-insensitive species the correlations among different spatial scales are consistent at some spatial scales but at other scales.The results also suggest that the results of azonal species at the local scale near their distribution boundaries can not be extrapolated simply to broader scales due to stronger responses to climate warming in those boundary regions.

DEM Production/Updating Based on Environmental Variables Modeling and Conflation of Data Sources

Availability of digital elevation models （DEMs） of a high quality is becoming more and more important in spatial studies. Standard methods for DEM creation use only intentionally acquired data sources. Two approaches which employ various types of data sets for DEM production are proposed： （1） Method of weighted sum of different data sources with morphological enhancement that conflates any additional data sources to principal DEM, and （2） DEM updating methods of modeling absolute and relative temporal changes, considering landslides, earthquakes, quarries, watererosion, building and highway constructions, etc. Spatial modeling of environmental variables concerning both approaches for （a） quality control of data sources, considering regions, （b） pre-processing of data sources, and （c） processing of the final DEM, have been applied. The variables are called rate of karst, morphologic roughness （modeled from slope, profile curvature and elevation）, characteristic features, rate of forestation, hydrological network, and rate of urbanization. Only the variables evidenced as significant were used in spatial modeling to generate homogeneous regions in spatial modeling a-c. The production process uses different regions to define high quality conflation of data sources to the final DEM. The methodology had been confirmed by case studies. The result is an overall high quality DEM with various well-known parameters.

Temporal variability of temperature-nitrate relationship in a coastal region

The inverse relationship between nitrate and temperature(N-T relationship) has been used to estimate new production from remotely sensed sea surface temperature at the regional or global scale of oceans. This study aimed to develop a time-series model of the N-T relationship from automated, continuous hourly observations over two years on the coast of Halifax, Canada. The model demonstrated time-series variability of the N-T relationship at a coastal station on the Nova Scotia Shelf, with adjusted R 2 =0.999 4 and RMSE=0.025 7. The maximum residual value was 0.077. The annual temperature variations described a sine curve, and daily, weekly, and monthly variations fluctuated within the normal ranges, controlled by the local climate. The annual variation of nitrate concentration formed nearly a sine curve. Heavy or longlasting rainfall increased nitrate concentration by 4 to 30-fold in 24 h, and then the increased nitrogen was quickly depleted by phytoplankton growth in 10 to 48 h. In general, biological activity was a key factor in causing nitrate concentration change, dependent mainly on seawater temperature. The power function of the N-T relationship observed in our study area could be used to quickly estimate sea surface nitrate concentration, in combination with temperature data obtained by remote sensing.

MODELING INTRAPERSONAL DEFORMATION SUBSPACE USING GMM FOR PALMPRINT IDENTIFICATION

In this paper, an efficient model of palmprint identification is presented based on subspace density estimation using Gaussian Mixture Model (GMM). While a few training samples are available for each person, we use intrapersonal palmprint deformations to train the global GMM instead of modeling GMMs for every class. To reduce the dimension of such variations while preserving density function of sample space, Principle Component Analysis (PCA) is used to find the principle differences and form the Intrapersonal Deformation Subspace (IDS). After training GMM using Expectation Maximization (EM) algorithm in IDS, a maximum likelihood strategy is carried out to identify a person. Experimental results demonstrate the advantage of our method compared with traditional PCA method and single Gaussian strategy.

Walter Benjamin： A New Positive Concept of Destruction

The author＇s proposal is to show how an apparently marginal text by Walter Benjamin, ＂The Destructive Character＂, can be taken as a hermeneutic model for understanding the connection of Benjamin＇s with history and tradition and for characterizing his thought as philosophy of threshold. It is the author＇s intention to show how the destructive character, on the threshold between destruction and construction, succeeds in remaining in that transition area that announces a change, showing its creative and conservative features; thus it appears as a figure of coexistence between stillness and movement, passage and interruption, and past and future. What the author wants to point out is that, in the description of this figure, Benjamin is driven by a cognitive demand, which tends to focalize the process of destruction itself rather than the result of the destruction process. The space in which the destructive character is acting is the space of the possibility of change and emancipation, and the space where it is possible to keep together past and future, and memory and redemption. Starting from these reflections, an important question arises about the problem of the relation between history and tradition, and culture and its memory： Neither the destruction of memory by a culture nor its canonization is meaningful. The concept of destructive character then represents a preferential point of observation for analyzing the articulations of Benjamin＇s thought. From this perspective, a decisive point of Benjamin＇s reflection seems to emerge on one side, strengthened in the permeation of Hebraic ancestors between destruction and salvation; on the other side, some of the concepts and images of the Berliner philosopher seem to show up in a new light.

An Extended Closed-loop Subspace Identification Method for Error-in-variables Systems

A closed-loop subspace identification method is proposed for industrial systems subject to noisy input-output observations, known as the error-in-variables (EIV) problem. Using the orthogonal projection approach to eliminate the noise influence, consistent estimation is guaranteed for the deterministic part of such a system. A strict proof is given for analyzing the rank condition for such orthogonal projection, in order to use the principal component analysis (PCA) based singular value decomposition (SVD) to derive the extended observability matrix and lower triangular Toeliptz matrix of the plant state-space model. In the result, the plant state matrices can be retrieved in a transparent manner from the above matrices. An illustrative example is shown to demonstrate the effectiveness and merits of the proposed subspace identification method.

Applications of variogram modeling to electrical resistivity data for the occurrence and distribution of saline groundwater in Domail Plain,northwestern Himalayan fold and thrust belt,Pakistan

This paper studies electrical resistivity dataset acquired for a groundwater study in the Domail Plain in the northwestern Himalayan section of Pakistan. Through a combination of geostatistical analysis,geophysical inversion and visualization techniques,it is possible to re-model and visualize the single dimension resistivity data into 2D and 3D space.The variogram models are utilized to extend the interpretation of the data and to distinguish individual lithologic units and the occurrence of saline water within the subsurface. The resistivity data has been calibrated with the lithological logs taken from the available boreholes. As such the alluvial system of the Domail Plain has formed during episodes of local tectonic activity with fluvial erosion and depositionyielding coarse sediments with high electrical resistivities near to the mountain ranges and finer sediments with medium to low electrical resistivities which tend to settle in the basin center. Thus a change is depositional setting happened from basin lacustrine environment to flash flooding during the Himalayan orogeny. The occurrence of rock salt in the northern mountains has imparted a great influence on the groundwater quality of the study area. The salt is dissolved by water which infiltrates into the subsurface through the water channels. Variogram aided gridding of resistivity data helps to identify the occurrence and distribution of saline water in the subsurface.

ANALYTICAL SOLUTION WITH SYSTEM METHOD FOR THE MODEL OF ROCK STRATA MOVEMENT AS COMPLEX LAMINATED PLATES

Starting from the basic equation of elastic mechanics, without any additional hypotheses of displacement or stress model, just introducing state space and state equation, the model of rock strata movement as complex laminated plates is presented. In addition, using displacement as a basic unknown quantity, the accurate analytical series solution for the problem of strata movement induced by extraction of horizontal seam is worked out when crosswise isotropic elastic layers are in sliding contact condition. A new approach is put forward to solve the complicated system of mining subsidence.

Spatio-temporal Distribution Pattern of Vegetation Net Primary Productivity and Its Response to Climate Change in Buryatiya Republic,Russia

As a critical index for monitoring the response of terrestrial ecosystems to climate change, net primary productivity （NPP） is widely used to reflect the vegetation productivity under natural conditions. However, little is known about vegetation NPP in Buryatiya Republic, Russia. To fully understand temporal and spatial variations in vegetation NPP and its response to climate change, satellite images of Moderate-resolution Imaging Spectroradiometer （MODIS） Normalized Difference Vegetation Index （NDVI） products in conjunction with ground-based observations of climatology were used to estimate vegetation NPP in Buryatiya Republic from 2000 to 2008. Observed values of vegetation NPP and MODIS NPP products were jointly utilized to verify simulations of vegetation NPP in this area. The Carnegie Ames Stanford Approach （CASA） model was improved by adjusting its parameters with bio-temperature instead of monthly average temperature in a sub-model of soil water content. Model reliability was tested with observed vegetation NPP. Variation in vegetation NPP, and its relationship with climatic factors, was analyzed in detail. Results showed that the modified CASA model can be used to estimate vegetation NPP in Buryatiya Republic, and that the accuracy of the model was generally high with a correlation coefficient of 0.91 （P0.01）. The vegetation average annual NPP was 542.45 g C m^-2 y^ -1 and the total NPP was estimated as 1.91E＋14 g C y^ -1 from 2000 to 2008. Monthly vegetation NPP increased significantly from April to July, decreased rapidly from August to November, and remained steady from December to March of the next year. Vegetation NPP increased with increasing longitude and decreased with increasing latitude. Different vegetation types differ in total NPP, with the highest NPP in forests, followed by forest-steppe, high mountain vegetation, steppe and meadow-swamp. Temperature and precipitation were the primary factors limiting vegetation NPP in this area.

Spatio-temporal variation of the wet-dry conditions from 1961 to 2015 in China

As an important part of the regional environment, the wet-dry climate condition is determined by precipitation and potential evapotranspiration （expressed as ETo）. Based on weather station data, this study first calculated ETo by using the FAO56 Penrnan-Monteith model. Then, the dryness index K （ratio of ETo to precipitation） was used to study the spatio-temporal variation of the wet-dry condition in China from 1961 to 2015; moreover, dominant climatic factors of the wet-dry condition change were discussed. The annual precipitation and ETo of the Qinling-Huaihe line were close to a balance （K≈1.0）. The annual precipitation in most areas exceeded the ETo in the south of this line and the east of Hengduan Mountains （K〈0.0）, where the climate is wet. Furthermore, the precipitation in the northwest inland areas of China, where the climate is dry, was markedly lower than ETo （K≥4.0）. The overall annual K of China fluctuated around the 55-year mean and its linear trend was not significant. However, a relatively wet period of about 10 yr （1987-1996） was recorded. The overall annual K of China showed strong cyclicality on the time scale of 3, 7-8, 11 and 26-28 yr, and regional differences of the annual K trends and cyclicality were large. The degrees of wetness in the Northwest China and western Qinghai-Tibet Plateau were substantially increased, whereas the degrees of dryness in the Yunnan-Guizhou Plateau, Sichuan Basin, and Loess Plateau were markedly increased. The linear trend of the annual K in most regions of China was not significant, and the annual K of most areas in China showed strong cyclicality on the 8-14 yr time scale. Precipitation was the dominant factor of wet-dry condition change in most areas, especially in North China, where the annual K change was highly correlated with precipitation.

Unified plastic modulus in the bounding surface plasticity model

A unified plastic modulus parameter for the bounding surface plasticity model is introduced in order to maintain the identical responses of modeling for both the two-dimensional and three-dimensional stress space with the same model parameters. Also discussed are the influences of the plastic modulus parameter on the stress-strain relationship and the plastic modulus. The model is more sensitive in modeling the stress strain responses when the plastic modulus parameter is small. The plastic modulus parameter has a great influence on the magnitude of the plastic modulus, especially at the initial loading stage. The plastic modulus asymptotically tends to zero at the end of loading.