规则二次曲面壳密度可变的全四边形网格剖分

本文提出了一种规则二次曲面壳上密度可变的全四边形网格剖分方法。利用二次曲面壳的标准参数方程，在二次曲面块Ｒ与参数平面的一个等参面Ｐ之间建立一种一对一映射关系。根据曲面壳上网格划分的密度要求，先在等参面Ｐ上剖分网格，然后将其精确映射到二次曲面Ｒ上。本方法消除了网格剖分时曲面上极点可能出现的奇异性，实现了二次曲面上的全四边形网格剖分，并可按网格划分的密度要求实现网格的疏密过渡。以球面，圆柱面为例，介绍     

用粒子群优化算法重构超二次曲面三维模型

针对在三维重构过程中用L—M（Levenberg—Marquardt）方法求解超二次曲面参数拟合问题的不足，提出了用粒子群优化算法来进行超二次曲面参数拟合的新方法．本文详细阐述了超二次曲面的三维表示特性，L—M算法拟合超二次曲面参数模型的分析，以及用粒子群优化算法拟合超二次曲面参数模型的原理、实现方法和实验结果．用粒子群优化算法对超二次曲面进行参数拟合，克服了L—M方法的缺陷，取了满意的效果．     

Polycube参数化自动构造

提出了Polycube参数化的自动构造技术.该算法首先对网格进行特征分解,然后用立方体组成的一些基本形体逼近分解得到的各部分网格区域,确定基本Polycube的顶点和边在区域上的对应顶点和路径,将各区域进一步分解为面片,从而在构造Polycube的同时完成对曲面的分片,最后再分片参数化并进行面片间的平滑,高效地实现了Polycube的自动参数化.该方法在很大程度上减少了Polycube构造过程中的人工干涉,使其能够在纹理映射等方面得到应用.     

区域分解型并行预条件的一种粗网格校正算法

区域分解是并行计算的基本手段之一,在稀疏线性方程组迭代求解时,对不完全分解等串行计算时很有效的预条件,经常采用区域分解的思想进行并行化。但区域分解的本质是利用局部解来近似全局解,从而必然存在较大误差,为此,提出一种粗网格校正算法,通过非重叠子区域浓缩,每个非重叠子区域浓缩为一个超结点,形成一个含全局信息且阶数等于子区域个数的小线性方程组,之后用其对原并行预条件进行校正。对块Jacobi型、经典加性Schwarz、以及因子组合型并行不完全分解预条件的实验表明,粗网格校正能有效改善收敛性并提高求解效率。     

一种子空间分解的MIMO-OFDM半盲信道估计方法

信道估计是MIMO-OFDM系统实现优良传输的一项重要环节。半盲信道估计算法是将MIMO-OFDM信道矩阵进行分解,分别利用未知数据和已知导频信息来完成信道估计。在利用未知数据进行估计时,提出一种利用频域子载波分组的子空间分解方法,不仅降低了计算复杂度,而且同时提高了信道估计的精度。利用已知的导频信息和未知数据估计出来的结果,可以求得最后的信道矩阵。相对于传统的频域子空间分解的半盲估计方法,算法可以减小计算复杂度80%,同时提高了估计精度平均1～2dB。仿真结果证明了算法具有良好的性能表现。     

基于信号功率谱密度分布的动态非均匀子带分解方法

针对频谱变化范围较大的宽带信号的子带分解问题,为了动态地调整子带的宽带与数量,合理地控制子带信号的自相关矩阵特征值扩散度,提高子带信号处理的性能和效率,在基于离散傅里叶变换(DFT)子带分解方法的基础上,提出了一种基于信号功率谱密度(PSD)的动态非均匀子带分解的新方法。对于给定序列,通过功率谱估计,确定子带数目和子带幅度范围,通过子带调制实现不同子带向零频处搬移,实现信号的分解。利用Matlab对子带信号特征值扩散度和信号重建性能进行了仿真。实验结果表明,与均匀子带分解相比,提出的方法直接利用PSD的分布信息实现非均匀子带分解,有效地控制了子带信号特征值扩散度在合理范围内的分布,并具有较好的重构性能。     

一种基于全域子空间分解挖掘的QoS准确预测方法

QoS的准确预测是评判和选择最佳Web服务的一种重要标准;传统的QoS预测方法采用时间平均值和各种参数简单加权的方法,无法对大量Web服务下的资源进行准确预测,预测结果模糊;提出一种基于全域子空间分解挖掘的QoS准确预测方法,即采用全域分析的思想对所有数据进行预处理,在此基础上,通过子空间分解的方法,在子空间中对数据进行分解分析,提取数据的深层次特征,然后将全域分析的结果与子空间分解分析的结果进行有效的数据融合,从而实现对所分析数据的准确预测;采用一组Web节点和拟定度量参数进行了预测实验,结果显示,基于全域子空间分解挖掘的QoS预测方法可以精确预测出渐变过程,结果准确,在QoS预测中具有广泛的应用价值。     

逆向工程中特征曲面的识别方法

提出了基于二次曲面拟合误差分析和几何属性分析的特征曲面识别算法。首先用一般二次曲面方程拟合测量数据并计算曲面的拟合误差，比较拟合误差与阈值的大小来识别自由曲面和一般二次曲面。然后计算测量点几何属并将它映射到高斯球和法曲率坐标上，运用统计方法识别不同分布模式，从而识别不同的特殊二次曲面类型。为了提高算法的效率和实用性，在计算测量点到拟合曲面的距离时，用近似距离代替精确的欧几里得距离。在识别特殊二次曲面时，优先识别平面。实例表明，该算法能有效识别自由曲面、平面、球面、柱面、锥面和过渡曲面等特征曲面。     

基于神经网络的机器人路径规划算法

机器人路径规划技术是机器人研究的一个重要领域.针对未知的全局环境,使机器路径最优化,利用机器人传感器网络建立可视区域,将整体任务分解为环境信息已知的一系列子任务,利用神经网络高速并行计算的优点,建立神经网络罚函数,提出一种实时性较高的变参数方法离散化求取罚函数的负梯度方向,控制机器人快速高效地完成子任务,从而驱使机器人到达目标点并进行仿真.仿真结果证明了复杂环境静态和动态目标指引下方法的有效性和实用性,特别适用于实时性要求高的场合.     

基于矢量场的二维区域全自动四边网格生成

为了实现任意二维几何模型的高质量分块结构四边形网格自动生成,提出一种基于矢量场的二维区域全自动分解方法.首先利用边界元法求解拉普拉斯型控制方程,获取一个反映模型边界几何特征、覆盖整个问题域的矢量场;然后结合矢量与标架的映射关系,将计算得到的矢量场转化为标架场;最后通过分析标架场的奇异结构将问题域分解成多个四边子区域,并在每个子区域利用映射法生成高质量的结构四边形网格.通过复杂区域的网格生成实例,验证了该方法的有效性和可靠性.     

Shape aware normal interpolation for curved surface shading from polyhedral approximation

Independent interpolation of local surface patches and local normal patches is an efficient way for fast rendering of smooth curved surfaces from rough polyhedral meshes. However, the independently interpolating normals may deviate greatly from the analytical normals of local interpolating surfaces, and the normal deviation may cause severe rendering defects when the surface is shaded using the interpolating normals. In this paper we propose two novel normal interpolation schemes along with interpolation of cubic Bézier triangles for rendering curved surfaces from rough triangular meshes. Firstly, the interpolating normal is computed by a Gregory normal patch to each Bézier triangle by a new definition of quadratic normal functions along cubic space curves. Secondly, the interpolating normal is obtained by blending side-vertex normal functions along side-vertex parametric curves of the interpolating Bézier surface. The normal patches by these two methods can not only interpolate given normals at vertices or boundaries of a triangle but also match the shape of the local interpolating surface very well. As a result, more realistic shading results are obtained by either of the two new normal interpolation schemes than by the traditional quadratic normal interpolation method for rendering rough triangular meshes.     

基于局部插值的双三次图像放大

以距离和边缘特征为约束,提出构造分片定义的双三次多项式曲面实现图像放大 的新方法,分为构造拟合曲面和修正曲面。以距离和边缘为约束构造对小邻域上像素拟合的二 次多项式采样曲面,所有二次多项式采样曲面加权组合生成分片定义的双三次多项式整体曲面。 由放大图像计算误差图像,由误差图像构造修正曲面的技术,进而提高放大图像精度和视觉效 果。为减少构造二次多项式的计算量,提出对二次多项式系数分类计算算法,能够实现对图像 任意倍数的放大。实验结果表明,该方法不仅提高了放大图像的峰值信噪比(PSNR)、结构相似 度(SSIM)数值精度,也提高了图像的视觉效果。     

非高斯随机粗糙表面计算机仿真的研究

在摩擦学、光学等工程领域中,对于表面粗糙度的研究,总是以生成的随机粗糙表面为研究对象,且大多数研究都是建立在高斯随机表面的基础上,而实际的工程表面大多是非高斯随机表面.为此提出了一种基于快速傅里叶变换(FFT)、Johnson转换系统和自相关函数等理论仿真生成非高斯随机粗糙表面的方法,它可以生成具有给定偏斜度和峰度的随机粗糙表面.为了说明该方法的可行性和正确性,给出了不同偏斜度、峰度和自相关长度下的计算机仿真结果.结果表明:在一定的条件下用该方法仿真生成的非高斯随机粗糙表面,其输入的随机表面的统计参数与输出的统计参数吻合较好.     

C1 Quadratic Interpolation over Arbitrary Point Sets

New formulas for generating smooth surfaces over arbitrarily spaced data points are developed. The formulas are based on quadratic polynomials for the construction of derivative continuous surfaces rather than on the cubic polynomials generally used. The technique is based on a subdivision procedure, dividing each triangle in a triangulation of the data points into six subtriangles and fitting a quadratic Bezier surface patch over each subtriangle. THe formulas require only function and first derivative values at the data points and are easily evaluated in terms of the Bezier coefficients. Since two-dimensional quadratic polynomials contain only six terms, while 10 terms are required to evaluate a cubic, the new procedure significantly improves the efficiency of algorithms for drawing surfaces in computer-aided geometric design.     

Quadratic curve and surface fitting via squared distance minimization

Quadratic curve and surface fitting to a set of data points are fundamental problems in reverse engineering and many other application areas. We develop the fitting methods for quadratic curves and surfaces based on the squared distance minimization technology. The basic idea of squared distance minimization for curve and surface fitting is first presented. Then we devise the corresponding squared distance term for each quadratic curve and surface, and minimize it to obtain its parameters. We repeat the squared distance minimization and update the parameters of the quadratic curve and surface by iterations until convergency. Consequently, the final fitting result is achieved. Experimental results demonstrate the effectiveness of the fitting method.     

Self-intersections of offsets of quadratic surfaces: Part II, implicit surfaces

The paper investigates self-intersections of offsets of implicit quadratic surfaces. The quadratic surfaces are the simplest curved objects, referred to as quadrics, and are widely used in mechanical design. In an earlier paper, we have investigated the self-intersections of offsets of explicit quadratic surfaces, such as elliptic paraboloid, hyperbolic paraboloid and parabolic cylinder, since not only are they used in mechanical design, but also any regular surface can be locally approximated by such explicit quadratic surfaces. In this paper, we investigate the rest of the quadrics whose offsets may degenerate, i.e. the implicit quadratic-surfaces (ellipsoid, hyperboloid, elliptic cone, elliptic cylinder and hyperbolic cylinder). We found that self-intersection curves of offsets of all the implicit quadratic surfaces are planar implicit conics and their corresponding curve on the progenitor surface can be expressed as the intersection curve between an ellipsoid, whose semi-axes are proportional to the offset distance, and the implicit quadratic surfaces themselves.     

三角网格上的代数曲面重建

提出一种用分片代数曲面构造三角曲面片的方法,利用具有公共边的2个三角形区域的4个顶点的函数值以及公共边2个端点的外法向量来构造一个二次曲面V(g)和一个截面V(h),其交V(g,h)即为2个三角曲面片的公共边界曲线.对每个已确定了边界条件的三角片内部进一步划分成3部分,每部分各自定义一个三次代数曲面.这3个三次代数曲面不仅在其交线处光滑拼接,而且分别沿三角形的边界与V(g)光滑拼接,从而构成一个具有GC1连续性的分片代数曲面.对于只属于一个三角片的边界留有一个自由度,可对曲面形状加以控制.     

Inferring Solids Composed of Linear and Quadratic Surfaces from Incomplete Three Views

This paper introduces an automatically reasoning system for inferring linear and quadratic surface solids from incomplete three views,The method is based on human reasoning .Initially,quadratic surface primitives are fecognized and extracted according to possibility computation.Then ,the data relative only to quadratic surface primitives are femoved from the given three views and polyhedrons are constructed based on the remaining data.The final solid is constructed by assembling all the generated quadratic surface primitives and the polyhedrons.The system has been implemented in C and OpenGL on Windows NT.The constructed solid can be observed at any angle by rotating it using the keyboard.Many examples are tested and the experimental results show that the system is very robust,and can accurately and effectively construct a suitable solid composed of linear and quadratic surfaces from incomplete three views.     

Recognition of quadratic surface of revolution using a robotic vision system

Reverse engineering using 3D scanners has been gaining increasing popularity. One challenging task that remains is to recognize the geometric feature from the cloud data scanned. In this study, a robotic vision system is used to recognize quadratic surfaces of revolution on an object.The top-view image of an object is used to detect the surface boundary by loop analysis technique. The boundary of a single surface is extracted according to the 2D loop of that surface. The robot then projects laser lines through the principal axes of the loop to get the sectional curves. The surface is recognized by a curve-fitting method based on the characteristics of these curves.This study provides a simple and faster method to detect the manufacture features on an object that contains quadratic surfaces. The data structure can be output in IGES format for re-design or rapid manufacture of the object.     

带形状参数的QT-Bézier曲线曲面的构造和应用

为了更加方便地表示和修改曲线曲面,提出了带形状参数的四次三角Bézier曲线曲面QTBézier的构造方法和应用。首先仿照Bézier曲线性质,构造了带形状参数的基函数,定义了带形状参数的QT-Bézier曲线曲面并研究了他们的一些主要性质,并就参数的选取做了一些分析。这种带形状参数的QT-Bézier曲线曲面是已有的一些曲线曲面的一般表达方法,如果选取一些特殊的参数,可以表示特殊的和已知的曲线曲面,还可以构造不同形状的旋转面。带形状参数的QT-Bézier曲线曲面可以很好地通过形状参数来调整曲线曲面的外形,而且能构造不同的旋转面,由于有额外的形状参数,更便于交互。