Catmull-Clark 细分曲面的误差分析

运用引入相邻点的方法和计算控制点的一阶差分的新的技术,研究Catmull Clark曲面细分过程的误差估计问题.证明了Catmull Clark曲面的控制网格按指数速率收敛于极限曲面;并给出关于Catmull Clark曲面误差估计的一个计算公式.本文的技术亦可适用于Doo Sabin曲面等细分曲面.     

Precise evaluation of uniform Doo-Sabin surfaces

Using eigenanalysis of uniform Doo-Sabin subdivision surfaces, parameterization approaches for two types of Doo-Sabin subdivision surfaces are developed, and evaluation methods for values and derivatives of Doo-Sabin surfaces at arbitrary parameters are presented by means of non-iterative approaches. Generalized basis functions for Doo-Sabin surfaces are also derived. Thus, the prejudice is disproved further that subdivision surfaces cannot be evaluated precisely.     

Defomable surface modeling based on dual subdivision

Based on dual Doo-Sabin subdivision and the corresponding parameterization, a modeling technique of deformable surfaces is presented in this paper. In the proposed model, all the dynamic parameters are computed in a unified way for both non-defective and defective subdivision matrices, and central differences are used to discretize the Lagrangian dynamics equation instead of backward differences. Moreover, a local scheme is developed to solve the dynamics equation approximately, thus the order of the linear equation is reduced greatly. Therefore, the proposed model is more efficient and faster than the existing dynamic models. It can be used for deformable surface design, interactive surface editing, medical imaging and simulation.     

Deformable surface modeling based on dual subdivision

Free formsurfaces ,especiallynon uniformratio nalB splines (NURBS) ,arewidelyusedincomputeraidedgeometricdesignandcomputer aideddesign/manufacturing .However ,usuallyasurfaceofarbi trarytopologycannotberepresentedbyasingleNURBS .Moreover ,itisverydifficulttosettheknotspacings,manipulatethecontrolverticesormodifytheweightstomeetsomedesignrequestsusingNURBS .Consequently ,subdivisionsurfacesappeartouniformlymodelanycomplexsurfacesofarbitrarytopology[1] .However ,interactivemethodsformodi …     

基于非四边形的逆向细分

在实际应用中,大部分细分后的网格是四边形网格与其它的三角形等非四边形网格的混合结构,本文根据这一特点,提出了逆向Doo-Sabin和逆向Midedge细分算法,并且通过实例证明该算法是有效可行的。     

Doo-Sabin曲面控制网格的收敛估计

Doo-Sabin细分曲面是定义在任意拓扑网格上的一种细分曲面的框架,它是双二次B样条曲面的一种推广.基于这个性质Doc-Sabin曲面被广泛应用于具有任意拓扑结构的复杂形体的造型.本文运用Doo-Sabin控制点的一阶差分技术来研究Doc-Sabin细分曲面控制网格的收敛问题.证明了Doc-Sabin曲面控制网格以指数速率收敛,并给出了一个计算估计公式.在此基础上可以给出Doo-Sabin曲面的误差估计的计算公式.     

细分曲面的NC刀轨生成算法及实现

提出了一种基于LOOP细分规则的NC精加工刀轨生成算法,该算法将细分曲面应用于CAD/CAM系统,适用于任意拓扑的三角网格模型;通过控制曲面等距误差来生成满足给定精度要求的NC刀轨.其核心思想是:首先计算LOOP细分曲面控制顶点的极限点和法矢量,然后从极限点开始,沿其法矢方向以球头刀的半径长度按照给定精度向外等距,获得等距曲面,最后在等距曲面上生成精加工数控刀轨.实例表明该算法稳定、高效、误差小.     

Reverse Loop subdivision with sharp features

0 Introduction By using a subdivision rule a control mesh will be refined withincreased detail . Asequence of subdivisions then produces a hierarchyof surfaces withincreasingreso- lution.Such a hierarchy allows the selection of a repre- sentation of surface at a particularlevel of detail ,whichis useful in many applications in both modeling and render- ing, such as view dependent rendering and progressive transmission. This paper introduces a newreverse Loop subdivision algorithm which can re…     

步进电机细分运动控制量的优化

步进电机细分控制量优化对提高细分精度、减小细分误差有一定的重要性和必要性。定义了时间、控制量及角位移三维空间，介绍了离散控制量的优化方法，并提出了离散量控制细分优化方案。该方法对任何种类的步进电机控制细分都有意义。     

基于外心对偶剖分的有限体积元法

考虑基于外心对偶剖分的椭圆型与抛物型方程的有限体积元法. 设原始三角形剖分的任意三角形单元的重心Q和外心C的距离满足|QC|=O(h²), 在此条件下, 证明了二阶椭圆型方程基于外心对偶剖分的有限体积元法的L²误差估计, 以及抛物型方程基于外心对偶剖分的半离散和全离散有限体积元格式L²和H¹误差估计.     

从有边界点云数据生成偏移细分曲面

提出一种根据给定精度从有边界点云数据生成偏移细分曲面新方法.它是基于有边界的Loop细分,运用细分的局部特性,通过循环修正、优化、自适应细分域曲面的控制网格,使域曲面不断逼近点云数据,通过对域曲面的标量偏移来表示待构曲面的细节特征.利用细分曲面的任意拓扑适应性、整体连续性,重构出具有细节特征的无需裁剪和拼接的待构物体偏移细分曲面.实例表明,该算法不仅具有稳定性,同时构造出的细分曲面还具有较高的重构精度,较好地满足工程实际需要.     

基于特征的结构分析有限元h型自适应模糊控制算法

基于特征分区的网格生成和模糊控制原理,提出了一种不严格依赖后验误差估计的自适应有限元模糊控制算法,采用基于奇异效应的误差估计式,使得自适应的过程能客观地控制误差的分布.     

基于细分的数字半色调技术

提出一种用二值图形设备显示多级灰度图象的方法。这种基于细分的数字半色调技术利用了人类视觉系统的生理特点，优先考虑在较大区域内源始图象与经半色调处理后图象间的误差。该技术可用图象灰度级的压缩。此外还对这一算法与有序颤抖法及误差扩散法进行了比较。     

基于MATLAB/Simulink仿真的永磁同步电机新型超螺旋二阶滑模转速控制

为进一步提高永磁同步电机转速外环控制性能,设计了一种新型超螺旋二阶滑模转速控制器。基于传统超螺旋算法,提出了一种新型超螺旋算法,包括设计了变指数取代传统算法中非线性项的常指数,采用分数阶代替整数阶,构造了变边界层非线性指数函数取代开关函数。采用新型超螺旋算法设计永磁同步电机转速控制器,并进行MATLAB/Simulink仿真,过程中使用鲸鱼算法优化控制参数。所得仿真结果为：相比指数趋近律与传统超螺旋算法,采用新型超螺旋算法设计的转速控制器的转速超调、转速受扰时的下降幅度、转速稳态跟踪误差最小。根据仿真结果,得出结论为：新型超螺旋二阶滑模转速控制器在系统收敛能力、抗扰能力及抖振抑制能力方面均优于指数趋近律滑模和传统超螺旋二阶滑模转速控制器。     

基于Loop细分的自适应细分曲面算法

首先研究了传统的Loop细分曲面算法,通过分析发现随着细分次数的增多细分算法中三角形网格片数增长过快。针对这一问题提出一种自适应细分曲面算法。算法根据相邻两个三角形面上的法向量的夹角,判断细分网格中较为光滑和非光滑的区域。实验结果表明,算法提高了数据处理速度,并且模型简单易实现。     

基于Catmull-Clark细分的曲面裁减运算研究

本文在Catmull-Clark细分曲面求交完成后,采用局部修改交点处的控制网格拓扑结构和局部修正控制网格顶点位置方法,给出了控制网格上的任意点在细分曲面位置上的计算推导,方便地实现了对Catmull-Clark细分曲面的裁剪运算。     

Semi-stationary subdivision operators in geometric modeling

Based on the view of operator, a novel uniform subdivision construction method is proposed for free form curve and surface design. This method can give an appropriate manner of parameter change in the subdivision iteration with fewer parameters and better shape control, such as building local revolving surfaces. The convergent property of order 2 subdivision surface is elegantly analyzed using computing techniques of matrix. This method is promised to be valuable in Computer Aided Design and computer graphics, due to the simplicity both in mathematical theory and practical implementation, the similarity to the B-spline curve and surface, G1 continuity, the affine invariability and local flexible control.     

二阶混沌系统基于指数终端滑模控制的投影滞后同步

研究了带有外部干扰变量和不确定项的二阶混沌系统的投影滞后同步问题。基于Lyapunov稳定性理论及指数型快速终端滑模策略的有限时间快速收敛性,提出一种新的滑模变结构控制,并设计相应的滑模曲面和控制律。所设计的控制器使误差系统状态变量能以较快的收敛速度在有限时间内进入指数终端滑模曲面并最终到达平衡点。数值仿真验证了该控制方法的有效性。     

Integral computation relating to rational curves based on approximate degree reduction

A new algorithm is presented for computing the volume of revolution, moment of area and centroid etc. , which are related to the integration of rational curves. In this algorithm rational curve with high degree is firstly approximated by those with lower degree through endpoint interpolation. And finally the closed form integration solution is derived for quadratic rational curve. The diminishing rate of the minimum norm of the perturbation vector needed by degree reduction is 0(2" ") when the interval is subdivided at the midpoint. Combining the subdivision with the degree reduction, we can obtain a faster convergence of integration approximation. A series of integral error bound functions which are fairly simple to compute are derived. The examples given in this paper show that this algorithm is a simple and time-saving method in computing with small tolerance.     

一种Bezier曲面模型嵌入水印的方法

介绍一种Bezier曲面片嵌入数字水印的新方法.对于由多段Bezier曲面片构成的曲面,每个Bezier曲面片通过四角的控制顶点,嵌入水印过程是将一个Bezier曲面片细分为两片.每一次细分在原始曲面片上将产生两个附加的顶点,通过曲面片细分使水印嵌入到曲面中.