遥感卫星图像几何粗校正的数据并行方法研究

主要研究星上遥感图像的实时几何粗校正问题．卫星遥感图像现在一般都大到上万个像素行和列，采用传统的单个处理器的串行方式在星上进行实时处理是难以满足应用要求的．提出了一种在一维PE阵列的SIMD计算机上采用基于处理元阵列平移的数据并行校正方法，并根据NASA的LANDSAT-1卫星的有关的参数，对该方法进行了详细讨论，给出了具体的实现方法．通过对复杂性和加速比的讨论，表明该方法在性能上比采用单个处理器的串行方法提高了N倍．     

基于切空间判别学习的流形降维算法

在基于图像集的流形降维问题中,许多算法的核心思想都是把一个高维的流形直接降到一个维数相对较低、同时具有的判别信息更加充分的流形上.投影度量学习（projection metric learning,简称PML）是一种Grassmann流形降维算法.该算法是基于投影度量,并且使用RCG（Riemannian conjugate gradient）算法优化目标函数,其在多个数据集上都取得了较好的实验结果,但是对于复杂的人脸数据集,如YTC其实验结果相对较差,只取得了66.69%的正确率.同时,RCG算法的时间效率较差.基于上述原因,提出了基于切空间判别学习的流形降维算法.该算法首先对于PML中的投影矩阵添加扰动,使其成为对称正定（symmetric positive definite,简称SPD）矩阵;然后,使用LEM（log-euclidean metric）将其映射到切空间中;最后,利用基于特征值分解的迭代优化算法构造判别函数,得到变换矩阵.对提算法在多个标准数据集上进行了实验验证,并取得了较好的实验结果,从而验证了该算法的有效性.     

A stochastic model for electrodeposition process with applications in filtering and boundary control

Electrodeposition is a complex partially observed mass-transfer process driven by several surface reactions without exact model. In this article, the process uncertainties are described by a finite number of Wiener processes in a stochastic model applied in the filtering and control problems. These problems are solved as a boundary observation-control problem based on a finite diffusion model with uncertainties in the domain interior and on the boundaries. A mixed boundary problem is considered on an interval with the Dirichlet data on one end (bulk solution) and Neumann data on the other end (cathode surface). The concentration of oxidising species in the domain interior is unattainable for observations but the flux on the boundary (electric current) can be measured with a limited accuracy (sensor error). The total flux for the main and side reactions is controlled by the current density on the cathode surface. The disturbing effect of the side reactions is modelled as a noise. The concentration of species is stabilised at the desired level near to the cathode surface with a relatively simple feedback control. The concentration on the boundary and in the domain is estimated as a conditionally Gaussian process in the course of filtering. The estimated conditional mean of concentration is solved from a stochastic partial differential equation in dependence on the covariance kernel. A relatively good quality of estimation and control is demonstrated in the process of simulation in the realistic conditions for a copper deposition process.     

NURBS曲线曲面插补算法基于S12的测试

传统的基于直线和圆弧的插补算法已经不能满足数控技术发展的需要,采用基于自由曲线的插补算法是大势所趋。目前关于NURBS曲线插补的研究大部分还停留在单段的研究上,没有突破传统数控技术的范围。在目前开放式数控技术日益重要的情况下,提出了基于整条曲线或者整个曲面的插补算法,并将该算法应用在S12单片机上进行测试,测试结果显示能够满足要求。     

Complex Valued Spectral Hermite Approximations for the Actively Mode-Locked Laser

We construct a numerical approximation of the governing equations of an actively mode-locked laser. The governing equation is complex valued and a novel scaling is employed that is designed to simplify the associated line integral in the complex plane. The resulting approximation is based on a set of shifted Hermite polynomials on an infinite line. Numerical comparisons are given with a finite difference scheme on a mapped domain as well as a finite element method on a truncated domain.     

基于活跃用户特征的论坛广告投放

网络论坛用户众多,广告投放市场大,而目前针对网络论坛的个性化广告却很少,这是因为论坛数据庞杂,信息更新快,实现广告的精准投放有困难.提出了一个基于活跃用户特征词分析的论坛广告个性化投放模型,在模型中提取论坛活跃用户,对活跃用户偏向性进行分析,从而以活跃用户为载体进行个性化广告投放.最后将这一模型在一热门论坛数据上进行了实验,实验结果表明该模型能实现论坛广告的个性化投放.     

Operations and implementation of complex objects

A model of a complex object is presented and a set of meaningful operations, both basic and advanced, on a single complex object and on a configuration of complex objects is defined. A set of requirements is presented for storage subsystems that support complex objects. Implementation of complex objects and operations on a single complex object are described, and a detailed performance analysis is provided which establishes the merit of complex objects. Finally, storage techniques are proposed for supporting advanced operations on a configuration of complex objects     

MEMS packaging process by film transfer using an anti-adhesive layer

A low cost and low temperature thin film packaging process based on the transfer of an electroplated Nickel 3D cap is proposed. This process is based on adhesion control of a thick molded cap Ni film on the carrier wafer by using a plasma deposited fluorocarbon film, on mechanical debonding and on adhesive bonding of the microcaps on the host wafer with BCB sealing rings. Mechanical characterizations show that the transferred microcaps have a high stiffness, a low stress and a high adhesion. Because this process is simple and only involves a low temperature (250°C) heating of the host wafer, it is highly versatile and suitable for the encapsulation of micro and nano devices, circuits and systems elaborated on a large range of substrate materials.     

Multi-level graph layout on the GPU

This paper presents a new algorithm for force directed graph layout on the GPU. The algorithm, whose goal is to compute layouts accurately and quickly, has two contributions. The first contribution is proposing a general multi-level scheme, which is based on spectral partitioning. The second contribution is computing the layout on the GPU. Since the GPU requires a data parallel programming model, the challenge is devising a mapping of a naturally unstructured graph into a well-partitioned structured one. This is done by computing a balanced partitioning of a general graph. This algorithm provides a general multi-level scheme, which has the potential to be used not only for computation on the GPU, but also on emerging multi-core architectures. The algorithm manages to compute high quality layouts of large graphs in a fraction of the time required by existing algorithms of similar quality. An application for visualization of the topologies of ISP (Internet Service Provider) networks is presented.     

一种基于信号-噪声模型的主元数目选择方法

主成分分析(PCA)是一种基于二阶统计的数据分析方法。在简要介绍PCA的基本理论后,重点分析了主元数目的选择问题。在基于累计贡献率的传统方法基础上,考虑实际中噪声对信号的影响,提出一种基于信号-噪声模型的主元数目选择新方法。实验证明,在一定的条件下,新方法比传统的累计贡献率方法更有效。     

Residual-based specification of a hidden random field included in a hierarchical spatial model

A method for specifying a hidden random field (HRF) included in a hierarchical spatial model is proposed. In hierarchical models of interest the first stage describes, conditional on a realization of the HRF, a response variable which is observable on a continuous spatial domain; the second stage models the HRF which reflects unobserved spatial heterogeneity. The question which is investigated is how can the HRF be modeled, i.e. specified. The method developed to address this question is based on residuals obtained when the base model, i.e. the hierarchical model in which the HRF is assumed constant, is fitted to data. It is shown that the residuals are linked with the HRF, and the link is used to specify the HRF. The method is applied to simulated data in order to assess its performance, and then to real data on radionuclide concentrations on Rongelap Island.     

Signal detection for bilinear systems

The injection operator used by McKean to construct Brownian paths on a Lie group is employed to formulate a class of signal detection problems on matrix Lie groups. The hypotheses that the signal is absent and present in the observation on a Lie group are described by a pair of bilinear matrix Ito equations. The injection operator is shown to be almost surely bijective and its inverse is constructed. This bijection enables us to do detection and estimation of processes taking values on a Lie group in a finite dimensional linear space-the Lie algebra. In fact, bilinear detection on Lie groups (“locally linear spaces”) is in essence equivalent to linear detection on linear spaces. A likelihood ratio formula for bilinear matrix Ito equations is thus derived as a function of up-dated observation.     

一种基于Linux工作站网络的分布并行计算的设计方法

基于工作站网络的分布并行计算是并行计算的一个重要发展方向。而Linux系统是目前有发展前途的支持并发开放的操作系统。文章提出了一种在Linux工作站网环境下,基于CORBA技术和多进程编程实现分布并行计算的程序设计方法,经过实例测试,表明该方法具有实用价值。     

A marker-less assembly stage recognition method based on corner feature

Fast and robust product assembly stage recognition is a key step in human–machine cooperative assembly. To solve the recognition problem of similar adjacent assembly stages, a marker-less assembly stage recognition method is proposed based on corner feature between an assembling product and a digital model. Considering the geometric features of mechanical products, a corner identification method is proposed based on the circumferential angle difference (CADF). Then a corner matching method based on distance constraint is studied for ICP registration to realize the point cloud registration between the product and digital model. Based on the registration relationship, a similarity algorithm based on proximity point proportion is used to calculate the similarity between models and the input assembling product. The model with the greatest similarity is taken as the stage recognition result. In experiments on four group assembling products, the average stage recognition accuracy is 96.15%, which indicates that the proposed method can solve the stage recognition problem. The corner identification method based on the CADF outperforms the Harris-3D corner detection method in the efficiency of assembly stage recognition.     

Systematic methods for the computation of the directional fieldsand singular points of fingerprints

The first subject of the paper is the estimation of a high resolution directional field of fingerprints. Traditional methods are discussed and a method, based on principal component analysis, is proposed. The method not only computes the direction in any pixel location, but its coherence as well. It is proven that this method provides exactly the same results as the "averaged square-gradient method" that is known from literature. Undoubtedly, the existence of a completely different equivalent solution increases the insight into the problem's nature. The second subject of the paper is singular point detection. A very efficient algorithm is proposed that extracts singular points from the high-resolution directional field. The algorithm is based on the Poincare index and provides a consistent binary decision that is not based on postprocessing steps like applying a threshold on a continuous resemblance measure for singular points. Furthermore, a method is presented to estimate the orientation of the extracted singular points. The accuracy of the methods is illustrated by experiments on a live-scanned fingerprint database     

三角形域上的超限插值方法

在三角形域上构造对边界曲线和跨界导数插值的三角曲面是计算机辅助几何设计和计算机图形学等领域中的基本问题.此类问题称为三角形域上的超限插值问题.对现有三角形域上的超限插值方法进行了综述,并对现有三角形域上的超限插值方法以具体实例进行了比较.最后讨论了现有三角形域上的超限插值方法中有待进一步解决的问题.     

Modeling and experimentation of a passive low frequency nanoforce sensor based on diamagnetic levitation

This paper is focused on the study of a new low frequency micro and nanoforce sensor based on diamagnetic levitation. The force sensitive part is a 10-cm long macroscopic capillary tube used as a levitating seismic mass. This tube presents a naturally stable equilibrium state with six degrees of freedom thanks to the combination of diamagnetic repulsive and magnetic attractive forces. It is only used as a one-direction force sensing device along its longitudinal axis. This force sensor is passive. The force measurement is based on the displacement of the capillary tube and in steady-state this displacement is proportional to the force. This sensor is characterized by an under-damped second-order linear force-displacement dynamic which remains linear on several hundred micrometers and can thus measure a wide range of microforces. Because of the magnetic springs configuration used, the capillary tube presents a horizontal mechanical stiffness that can be adjusted between 0.01 and 0.03 N/m (similar to the stiffness of a thin AFM cantilever). The measurement range typically varies between ±50 μN. Bandwidth is 4 Hz. The resolution depends on the sensor used to measure the capillary tube displacement and on noises induced by environmental conditions (ground and air vibrations). The resolution typically reached with a STIL confocal chromatic sensor is 5 nN inside a test chamber located on a anti-vibration table. This study is illustrated by a pull-off force measurement.     

A content-based goods image recommendation system

The information of e-commerce images varies and different users may focus on different contents of the same image for different purpose. So the research on recommendation by computers is becoming more and more important. But retrieval based only on keywords obviously falls short for massive numbers of resource images. In this paper, we focus on a recommendation system of goods images based on image content. Goods images have a relatively homogenous background and have a wide range of applications. The recommendation consists of three stages. First, the image is pre-processed by removing the background. Second, a weighted representation model is proposed to represent the image. The separated features are extracted and normalized, and then the weights of each feature are computed based on the samples browsed by the users. Third, a feature indexing scheme is put forward based on the proposed representation. A binary-tree is used for the indexing, and a binary-tree updating algorithm is also given. Finally, the recommended images are given by a features combination searching scheme. Experimental results on a real goods image database show that our algorithm can achieve high accuracy in recommending similar goods images with high speed.     

Minimum Projection Method for nonsmooth control Lyapunov function design on general manifolds

Asymptotic stabilization on noncontractible manifolds is known as a difficult control problem. On the other hand, an important fact is every control system that is globally asymptotically stabilizable at a desired equilibrium must have nonsmooth control Lyapunov functions.This paper considers the problem of construction of nonsmooth control Lyapunov functions on general manifolds, and we propose a nonsmooth control Lyapunov function design method called the ‘Minimum Projection Method’. The proposed method considers a simple-structured smooth manifold associated with the original manifold by a surjective immersion, and then a control Lyapunov function defined on the simple-structured manifold is projected to the original manifold. A function on the original manifold is thus obtained.In this paper, we prove that the control system on another manifold associated with a surjective immersion is determined uniquely, and the resulting function by the proposed method is a nonsmooth control Lyapunov function on the original manifold. The effectiveness of the proposed method is confirmed by examples.     

基于状态-动作图测地高斯基的策略迭代强化学习

在策略迭代强化学习中,基函数构造是影响动作值函数逼近精度的一个重要因素.为了给动作值函数逼近提供合适的基函数,提出一种基于状态-动作图测地高斯基的策略迭代强化学习方法.首先,根据离策略方法建立马尔可夫决策过程的状态-动作图论描述;然后,在状态-动作图上定义测地高斯核函数,利用基于近似线性相关的核稀疏方法自动选择测地高斯...