基于法向分类的三维模型形状分布检索算法

给出一种用三个形状分布图来描述三维模型的检索算法.在模型表面任意取点,并记录每个点的法向量.将任意两点连线形成线段,计算该线段的欧几里得距离,再分别计算两端点的法向与该线段的夹角.根据两个夹角的值,将所得线段分成三个集合,并分别构造三个集合的形状分布曲线,通过对模型间三条形状分布曲线的比较,得出两个模型的相似程度,从而实现模型的相似性检索.实验表明,该算法能够较好地实现三维模型检索,检索结果比传统的形状分布算法有较大的改进.     

一种距离特征融合的机械三维模型检索方法

三维模型检索技术可提高产品设计效率,减少设计时间。针对形状分布算法(D2)只适用于全局特征检索而对局部特征描述不足的缺点,文中提出一种融合局部特征和全局特征的三维模型检索方法。首先,运用蒙特卡洛算法在模型表面随机取点,计算随机点对间的D2距离和协方差距离,然后联合统计两种特征距离的分布概率,形成距离分布矩阵。最后采用Manhattan距离度量距离分布矩阵的相似性。实验结果表明,该算法可实现机械三维模型检索,检索性能明显优于两种经典算法。     

机械零件模型局部形状相似性检索算法研究

提出利用尺度不变特征来描述三维模型的局部形状信息,设计出一种灰度体素模型上的基于三维SIFT算法的局部形状特征提取方法;得到三维模型的尺度不变形状特征后,采用BoF方法构建模型的形状特征向量,配合相应的距离函数实现三维模型"部分-整体"的匹配和检索。实验表明该方法具有比较好的检索性能。     

一种三维CAD模型自动语义标注算法

针对三维CAD模型检索中的语义鸿沟问题,提出一种基于局部形状分布及语义概率统计的三维CAD模型自动语义标注算法.采用基于局部形状分布的多尺度特征提取方法获取CAD模型的形状信息,并计算不同模型之间的形状相似度;根据模型样本库中已知的语义分类信息,构建一个基于概率的标注框架对CAD模型进行语义标注,以建立模型形状信息和语义信息之间的联系.实验结果表明,该算法能够有效提高三维CAD模型检索的准确率,检索性能优于仅使用形状信息时的检索结果.     

一种保持特征的网格简化算法

针对网格简化算法中出现的细节特征丢失、简化结果过于均匀以及计算复杂等问题,从原始模型的几何特征出发,对待折叠三角形与其邻域内三角形的形状与位置关系进行研究,对三角形折叠点到其三个顶点的偏离程度进行总结,提出了一种基于三角形折叠的保持模型特征的网格简化算法。该方法根据目标三角形各顶点邻域三角形的不同,为其3个顶点分配相应的权值,由各顶点权值的大小最终确定三角形折叠点坐标,然后以二次误差测度为度量标准计算出每个三角形的折叠代价,再结合三角形局部面积比、局部区域不平度以更好地控制简化结果。实验结果表明:该方法能够降低计算的复杂度、提高简化速度,使模型的特征区域和非特征区域均保持一定的简化率,并较好地保持了原始模型的细节特征。     

基于形态学的牙齿模型交互分割

精确地分离出单颗牙齿是计算机辅助正畸设计的重要步骤,由于牙齿有不同的形状并且不同个体之间也有很大的差异,所以分离出单颗牙齿比较困难.介绍了一种从三维网格牙颌模型上精确分割单颗牙齿的方法.首先,计算三角网格顶点的曲率,根据顶点曲率值筛选出特征区域;由用户封闭特征区域,之后对特征区域进行形态学操作得到特征区域的骨架线;依据获得的骨架线可以把牙齿精确地分离出来.     

集成形状比较技术的产品实例检索研究

为有效地缩短新产品的开发周期,提出了集成形状比较技术的产品实例检索的框架.根据形状分布方法,对零部件三维模型进行相似性度量,采用模糊C均值聚类算法对已有产品实例的三维模型进行聚类.三维模型的类别形成高维的欧式空间,从而提出采用向量空间模型表达产品实例.给出了实例库建立和实例检索的过程.实验研究表明,该方法应用到具体设计领域,可以提高基于实例设计的自动化程度和检索的准确度.     

基于样例的建模方法

提出一种新的基于样例的几何造型方法,该方法利用三维模型之间的剪切/粘贴操作从已有的两个或多个三维几何模型构造出新的几何模型.设计并实现了一种基于隐式曲面布尔运算的网格模型剪切/粘贴算法,该算法首先将需要的部分网格从源模型上剪切下来,并将其配准到目标模型上;然后将两网格模型转化成点模型表示,并将点模型转化成径向基函数(RBF)的隐函数表示;再对两隐函数进行布尔运算;最后将布尔运算生的隐函数曲面进行多边形化,得到最终的网格模型.该算法的布尔运算不是直接对网格模型进行操作,避免了耗时的面线、面面求交计算.试验结果表明,该算法具有很好的网格融合效果,可用于影视动画中构造特殊夸张造型.     

融合通用形变模型信息的面部三维重建

基于图像对的立体重建是用于获取人脸三维信息的通用方法,但根据图像数据和重建算法所得到的三维重建结果存在各种误差,本文对通用形变模型进行改进并与三维立体重建融合以得到更精确的重建结果。首先使用Max-Margin对象检测算法来获取面部边界框,其中回归树集合法能直接从像素强度的稀疏子集识别面部特征点。然后通过PCA颜色模型生成形状和颜色的三维面部统计模型,利用ISOMAP算法将三维网格转换为二维表面并提取纹理信息,得到面部模型。最后在源网格上进行两步非刚性表面配准的变形过程:先通过对源网格进行二次采样来选择少量网格点来表示源的全局变化,并选取径向基函数(RBF)进行非刚性全局变形;再对源顶点进行Procrustes分析获得非刚性变换,再通过加权方案来进行k-近邻变换,得到平滑的局部变形。将单图像重建的面部模型,立体重建的面部模型和本文的面部变形模型与高质量扫描云图进行对齐比较,得到面部变形模型的3个RMS值分别为2.795 2,2.102 8和2.153 4,相比于其他模型,面部变形模型更接近高质量扫描云图,即与原图像一致性更高,误差更小。面部变形模型的定性和定量分析表明,立体重建与人脸一般形状信息的组合在几何信息的表达上优于基于通用模型的单个图像重建以及未考虑通用模型的立体重建。     

基于机械结构形态的三维CAD模型检索

针对三维CAD模型检索相似度度量特征信息不完备问题,提出基于机械结构形态的三维模型检索方法.通过分析机械结构形态的基本特征,构建基本体形态位置相似比较的黄金分割相似系数,研究基本体形态姿态相似比较的方向向量、法向向量比较规则,并定义曲线轮廓值指标计算基本体形态的轮廓相似度.综合上述3个比较指标及坐标系变换相似度,提出基于机械结构形态的三维CAD模型粗检索方法.并通过基本体转换模型特征相似度计算方法的研究,提出三维CAD模型精检索方法.通过实例验证结果表明,所提检索算法有效地实现了模型外形特征、几何尺寸、拓扑关系等信息的比较,提高了检索结果的准确性,同时实现了以机械结构形态为查询对象的三维模型检索.     

An adaptive mesh model for 3D reconstruction from unorganized data points

Reconstructing exact topology mesh from data points is one of the most important tasks in the fields of industrial CAD/CAE/CAM, computational vision and reverse engineering. In this paper, a deflation algorithm that integrates an adaptive mesh and physical constraint model is presented for the 3D reconstruction of geometric-closed shape (genus 0) from unorganized data points. First, an initial mesh is formed using the Delaunay algorithm. Second, an asymptotic deforming performance is accomplished to deflate initial mesh towards the local concave boundary step-by-step. In this phase, a physical constraint model of coupled particle systems based on particle dynamics and Newtonian law of motion is constructed, and the model dynamically controls mesh deformation as a behavior constraint. To guarantee that the resultant mesh is homeomorphous to the original surface of data points, a continuously deforming mechanism, visibility cone and collision-detecting criterions are designed. At last, experimental results in reverse engineering which supports the usefulness of this method for reconstruction .     

复杂曲面特征线自动提取技术研究

复杂曲面零件的几何模型重构是逆向工程的研究重点之一,由零件表面的数字化数据提取特征线是构造几何模型的重要步骤。针对由CMM获取的三维数据“点云”,提出了基于扫描线的三维分层式方法实现特征线的自动提取,提出了基于局部增量网格扩张的三维散乱数据三角剖分算法,实现任意曲面尖锐棱线的自动提取。实践说明,通过该算法得到的复杂曲面的边界线能够满足模型重建的工程需求。     

Efficiency of grid representation and its algorithms for areal 3D scan data

This paper describes the efficiency of a grid representation for an areal 3D scan data and the algorithms for managing measurement data captured by areal 3D scanners. Due to the measurement principles of areal 3D scanners, a measurement point is generated for each pixel of the imaging sensor inside the 3D scanner. Therefore, when the measurement points are perspectively projected on the image plane of the imaging sensor, each point has one-to-one correspondence to the imaging elements of the sensor that has a regular grid structure. By using this property, measurement points are represented by their depth values in a grid representation model. Compared to the conventional representation model, such as triangular mesh and cloud of points, the grid representation uses less memory and allows efficient algorithms for processing the measurement data captured by areal 3D scanners. This paper was recommended for publication in revised form by Associate Editor Soon Hung Han Minho Chang is a Professor at the department of mechanical engineering at Korea University in Seoul Korea. He received a PhD degree in Mechanical Engineering from MIT in 1996. He worked for Korea Institute of Science and Technology. His research interests include mechanical design, three-dimensional measurement, and CAD. Yun Chan Chung is a Professor in the department of die and mold engineering at Seoul National University of Technology, Korea. He worked for Cubictek and DaimlerChrysler, developing CAD/CAM systems mainly in die and mold making. He received PhD in Industrial Engineering from KAIST in 1996. His research interests include digital manufacturing, tool-path generation and verification, and software engineering.     

三角网格牙颌模型上的牙弓线半自动探测方法

介绍了一种在三维三角网格牙颌模型上半自动探测牙弓线的方法。首先使用三维激光测量系统对牙颌石膏模型进行测量获得数字模型;其次在网格模型上交互拾取4个参照点以确定咬合平面;接着使用顶点法矢和咬合平面作为参考,筛选出初始特征点并拟合出初始牙弓线;以初始牙弓线作为新的参考依据筛选出最终特征点,拟合最终参照点得到最终牙弓线。使用多副牙颌模型对该方法进行了测试,验证了该方法稳定、有效。     

Tool path generation by offsetting curves on polyhedral surfaces based on mesh flattening

Polyhedral surfaces are used as representation model for CAM and process planning purposes because of its simplicity for data exchange and geometric computation. However, there are few tool path planning strategies for such surfaces but isoplanar method. This paper presents a contour offset approach to tool path generation for three-axis ball-end milling of polyhedral surfaces, based on a novel method for offsetting curves on polyhedral surfaces. One of its salient features is to reduce the task of removing complex interfering of offsets from 3D physical surfaces to 2D plane by flattening mesh surfaces and avoid costly 3D Boolean set operations and relatively expensive distance calculation. Moreover, in practical implement, the procedures of calculating offset points and removing interfering loops are merged and carried out simultaneously results in an efficient tool path generation method. Empirical examples illustrate the feasibility of the proposed method.     

一种散乱数据的三角剖分新算法

根据逆向工程中散乱数据点规模越来越大的趋势,为缩小剖分时搜寻和遍历数据点的空间范围,提高算法效率,提出了一种大规模散乱数据的空间划分方法及相应的数据结构和编码方案.同时,提出了外连剖分和内连剖分的概念,给出了基于局部增量网格扩张的3维散乱数据点的空间直接三角剖分算法.该算法的总体计算复杂度为O（N）,与三角剖分的典型算法相比,有效降低了其时间复杂度,提高了剖分后网格的质量.     

基于离散点云直接映射STL文件的建模技术

随着计算机技术的发展,反求工程技术和快速原型技术的结合日趋紧密,纯复制的反求工程已不再适应市场发展的需要,在产品的设计和制造阶段都需要具备创新环节的反求工程技术的支持。提出综合优化的方法：对扫描后获取的离散扫描点进行平面的三角剖分及环域间的三角剖分,生成STL文件格式,并结合环域划分风格的特点,使用最小权和最小内角最大化相融合的方法,实现二维优化：当二维坐标释放到空间后会破坏三角形形状,影响模型的质量,用增加点的方法,得到三维优化后的原产品模型的STL文件格式。在布尔运算的基础上,通过对反求模型的三角网格进行任意叠加和任意截取,产生新产品的STL文件格式;对新产品进行总体形状控制、过渡圆角处理等空间优化工作,最终实现新产品创新,缩短了产品的开发周期,降低了开发成本。     

基于Gambit构建翼型网格的问题

在对前处理软件 Gamibt 构建二维绕流网格时,对美国国家航空咨询委员会（national adviso-ry committee for aeronautics,简称 NACA）翼型库得到的翼型数据进行了一定的修改,解决了某些点的取舍等问题,并在 Gamibt 当中成功建立 C 型网格,完成了网格质量检查。     

Mesh smoothing method based on local wave analysis

As the mesh models usually contain noise data,it is necessary to eliminate the noises and smooth the mesh.But existed methods always lose geometric features during the smoothing process.Hence,the noise is considered as a kind of random signal with high frequency,and then the mesh model smoothing is operated with signal processing theory.Local wave analysis is used to deal with geometric signal,and then a novel mesh smoothing method based on the local wave is proposed.The proposed method includes following steps:Firstly,analyze the principle of local wave decomposition for 1D signal,and expand it to 2D signal and 3D spherical surface signal processing;Secondly,map the mesh to the spherical surface with parameterization,resample the spherical mesh and decompose the spherical signals by local wave analysis;Thirdly,propose the coordinate smoothing and radical radius smoothing methods,the former filters the mesh points’ coordinates by local wave,and the latter filters the radical radius from their geometric center to mesh points by local wave;Finally,remove the high-frequency component of spherical signal,and obtain the smooth mesh model with inversely mapping from the spherical signal.Several mesh models with Gaussian noise are processed by local wave based method and other compared methods.The results show that local wave based method can obtain better smoothing performance,and reserve more original geometric features at the same time.     

Complete 3D surface reconstruction from unstructured point cloud

In this study, a complete 3D surface reconstruction method is proposed based on the concept that the vertices of surface model can be completely matched to the unstructured point cloud. In order to generate the initial mesh model from the point cloud, the mesh subdivision of bounding box and shrink-wrapping algorithm are introduced. The control mesh model for well representing the topology of point cloud is derived from the initial mesh model by using the mesh simplification technique based on the original QEM algorithm, and the parametric surface model for approximately representing the geometry of point cloud is derived by applying the local subdivision surface fitting scheme on the control mesh model. And, to reconstruct the complete matching surface model, the insertion of isolated points on the parametric surface model and the mesh optimization are carried out. Especially, the fast 3D surface reconstruction is realized by introducing the voxel-based nearest-point search algorithm, and the simulation results reveal the availability of the proposed surface reconstruction method.