一种自由曲面点云边界的快速直接提取方法

点云边界不仅作为表达曲面的重要几何特征，而且作为求解曲面的定义域，对重建曲面模型的品质和精度起着重要的作用。本文以战斗机座舱盖玻璃罩外壳的点云数据为例论述了一种自由曲面点云数据的边界直接提取方法，该方法通过设定分隔截面和计算每个截面上边界点，然后连接各个边界点就得到曲面的边界。该方法不仅具有较高的运算效率，达到了系统的实时性要求，而且能够比较精确地表达曲面的边界特征。     

车身外覆盖件设计中点云重构A级曲面技术

A级曲面重构是车身外覆盖件设计的关键环节,针对传统方法重构复杂自由曲面效率低下、灵活性差、曲面质量不高的状况,提出一种基于NURBS面片的A级曲面重构方法,曲面调整灵活、直观,有效提高了A级曲面重构的效率和质量.     

截面数据的B样条曲面重建研究

针对截面数据 ,研究其B样条曲面重建问题。给定节点矢量、依据最小二乘方法生成初始拟合曲面 ;为了提高拟合精度 ,采用迭代最近点方法优化修正数据点的参数 ,进而生成符合精度要求的拟合曲面。实际算例表明该算法简单、实用、可靠     

Shape reconstruction, shape manipulation, and direct generation of input data from point clouds for rapid prototyping

Shape reconstruction from point clouds has received considerable attention in recent years on account of its ability to directly integrate reverse engineering with rapid prototyping. The primary objective of this study is to develop an integrated system that enables one to generate input data for rapid prototyping by constructing complete shape models from point clouds obtained with various measuring devices, including laser scanners, digitizers, and coordinate-measuring machines. We first present a novel approach to reconstructing a shape from point clouds based on implicit surface interpolation combined with domain decomposition. We then propose various related algorithms for generating input data for rapid prototyping, ranging from shape manipulation to complete solid generation. The validity of this new technique is demonstrated for a variety of point clouds with differing degrees of complexity.     

用于增材制造的点云直接分层算法

针对三维点云分层问题,将连续层片构造过程解释为对运动曲线的追踪,提出一种非监督神经网络用于从散乱点中增量式重构各层曲线.首先用分层平面对点云重采样得到二维样本点,然后逐个输入网络,使其动态增长、删除、移动、改变连接,从而抽取样本的局部主成分和拓扑结构.网络在当前层学习收敛后,其结构表达了一组多义线,该组多义线同时作为下...     

基于Level Set方法的点云建模技术研究

描述一种利用偏微分方程(PDE)和变分法进行点采样几何模型重建算法.把符号距离函数与任意函数的度量函数作为内部能量,根据能量约束最小化条件,通过Level Set方法求解曲面的梯度流.使初始曲面随时间产生演化变形,直至逼近目标模型,完全消除重新初始化过程.实验表明,通过Level set方法能够把初始曲面快速自动演化变形到目标模型,对任意拓扑结构和带噪声的点云具有很强的适应能力.     

散乱点云谷脊特征提取

利用微分"化曲为直"的本质特性和离散计算方法,提出了一种基于局部重建的散乱点云谷脊特征提取算法。首先,利用离散Laplacian算子对点进行增强,通过阈值过滤标记潜在谷脊点。然后,在每个潜在谷脊点的局部邻域内构建紧附于潜在曲面、能反映该点局部几何特征信息的三角网格。最后,根据Weingarten映射的性质,估算潜在谷脊点的主曲率和主方向;将邻域大小作为尺度参数,利用简单直观的离散计算方法及线性插值方法,多尺度地判定一点是否为主方向上的曲率极值点,从而提取谷脊特征。实验结果表明:当点云规模为10 375个,谷脊点规模为1 129个时,执行时间仅为97.39ms;当点云规模达327 853个,谷脊点规模达到105 482个时,执行时间为3 956.12ms。该方法简单、稳定,避免了传统的利用拟合曲面再逼近微分量方法中由于曲面拟合带来的高时间代价,能快速有效地提取散乱点的云谷脊特征。     

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.     

基于自由曲面点云的RPM自适应分层研究

针对由激光三维扫描设备获取的点云数据及其后续RPM操作,提出基于栅带的自由曲面点云自适应分层方法.以栅带为基元,将点云投影栅带化,从中获取特征点集,结合相邻栅带距离最小值点找寻构建层间轮廓.定义投影点云栅带面积比为表面误差,建立其与堆层厚度的关系,通过表面误差逼近定值最终确定堆层厚度.经实例验证,此方法得到的分层模型充分满足RPM的要求.     

Rapid surface reconstruction from a point cloud using the least-squares projection

A new approach for the rapid and robust surface reconstruction from a point cloud is presented based on the distance field and the least-squares projection (LSP) algorithm. This novel approach works directly on the point cloud without any explicit or implicit surface reconstruction procedure. First, a coarse base polygonal model was created directly from the distance field for the given point cloud through the iso-surface extraction. After acquiring a rough base polygonal model, we obtain a quality polygonal model through the iterative refinement and least-squares projection which projects current working polygonal model onto the point cloud in a least-squares sense. The main contribution of this work is the robust and fast surface reconstruction from randomly scattered 3D points only without any further information. We demonstrate the validity and efficiency of this new approach through a number of application examples.     

Recommended slicing positions for adaptive direct slicing by image processing technique

Build time and accuracy are two contradicting issues that have been a major concern in rapid prototyping, and have led to the development of many slicing approaches including those applying adaptive slicing, direct slicing, and adaptive direct slicing concepts. Presented in this paper is an approach for adaptive direct slicing that applies image processing technique to determine appropriate thickness for each sliced layer and to recommend slicing positions on a 3D CAD model. Two orthogonal views of a model are captured and converted to be edge images before being analyzed, and based on the surface complexity on the two edge images, slicing positions are recommended. These positions are passed to the CAD software for slicing activities. This adaptive direct slicing approach has been implemented on LabVIEW platform and compared with uniform direct slicing approach and uniform cusp height approach. The results show that this slicing approach improved slicing performance by reducing the number of layer which has a direct impact on build time while maintaining surface quality at the same level as the thin uniform direct slicing. Since its inputs are the images of a CAD model instead of the model itself, this adaptive direct slicing supports any CAD software.     

利用邻域方向分布的机载激光雷达点云建筑物外轮廓提取

机载激光雷达点云密度分布不均、建筑物形状不规则且复杂多样等多种因素,导致现有建筑物轮廓提取方法存在参数难以设置、适应性较差等问题,为此本文提出一种利用邻域方向分布的机载激光雷达点云建筑物外轮廓提取方法.首先采用固定邻域点数分析各点的邻域方向分布,以获取不同邻域方向间的夹角,根据建筑物外轮廓点的特点,定义潜在轮廓点并将其...     

截面测量数据的B样条曲面拟合

针对激光扫描仪测量得到的截面型数据,提出了对截面线重新采样的方法,分析了重采样数据拟合曲线的误差,及平方误差随重采样比例变化的规律,从而通过控制误差来确定重采样数目.构造插值于截面数据点的非均匀B样条,再对其进行相同数目的采样,得到了呈矩形分布的点阵,采用双向平均积累弦长参数化方法确定、方向节点矢量,用插值的方法进行曲面拟合.此种方法不仅避免了对轮廓线进行相容性的处理,而且减少了模型数据量,提高了曲面重构的效率.     

The development of a technique for 3D complex surface reconstruction from unorganized point cloud

In this paper, we propose a new neural network based on our two-level adaptive hierarchical clustering algorithm. The algorithm is to manage unorganized points, so that the triangular mesh models can be correctly obtained by applying the triangular mesh creation algorithm. We also develop adaptive self-flipping triangle edges to improve triangular mesh structure. Only one parameter, the maximal edge length of triangle, is needed in the neural network. The proposed two-level consists of the first level for clustering the cloud of points that has same order of the maximal edge length into a same cluster and the second level for generating triangular surface model or drape surfaces over the points of the same cluster. The normal vector for the generated triangular 3D surface model can be obtained from the second level. This helps to generate the STL file or stereolithography format. From the experimental results it can be shown that the proposed method is very effective for clustering unorganized point clouds for generating a triangular mesh of complex surfaces.     

Machine path generation using direct slicing from design-by-feature solid model for rapid prototyping

This paper explains a new machine path generating system that its output is compatible with different rapid prototyping processes. The basis of this system is direct slicing from design-by-feature solid model. Slicing a computer-aided design (CAD) model through intersecting the model with the XY-plane at each Z increment is a well-known method of path generation. Slicing a CAD model is currently conducted through stereolithography (STL) file slicing, direct slicing, and additive direct slicing. A direct slicing approach inside a design-by-feature solid modeler is proposed. Autodesk Inventor solid modeler, as a design-by-feature solid modeler, is used for 3D solid modeling. The proposed system is implemented by Visual Basic codes inside Inventor. In this approach, first protrusion and subtractive features that form a model are extracted. Then, the intersection of each feature and the XY-plane is identified. Then, the internal and external loops are found. Depending on the specific rapid prototyping (RP) process requirements, internal or external hatch are also computed, respectively. Finally, a continuous path in required format is generated. The system reported in this paper has been successfully tested on several complex 3D models created in Inventor. The system offers customized output for different RP processes that need external or internal hatch pattern. The proposed approach for generating RP machine path through feature recognition inside design-by-feature solid modeler overcomes with the problems that are caused by imperfect STL files. Also, this system is capable of generating code compatible with major rapid prototyping processes.     

采用八叉树体素生长的点云平面提取

提出了一种高效的基于八叉树体素自适应生成与体素分层次生长的平面提取方法,其主要思路为采用体素信息统计的方式进行相关阈值参数的自动选定,以及基于体素的生长替代基于点的生长进行平面提取。首先,对点云进行八叉树初始剖分并计算其几何属性信息(包括法矢、特征值以及维度特征描述符等);然后,通过统计得到细分终止条件,并对初始八叉树进行进一步自适应剖分,得到一系列非均匀八叉树体素;最后,在体素层面进行区域生长阈值的统计与体素的分层次生长,进行点云平面的精细提取。利用4种不同类型的点云数据对本文算法进行了测试。实验结果显示:精度和召回率可以达到95%以上,表明本文算法对数据质量不敏感,可以自动适应不同平台采集的、不同分布密度和不同数据质量的激光点云,并且高效地得到精细的点云平面提取结果。     

曲面重构中散乱点云数据曲率估算算法的研究

获取测量点云数据的几何特征信息是曲面重构的基础,估算数据点方向矢量和曲率是点云数据处理中必须面对的问题。这里针对散乱测量数据点云,以局部数据点协方差矩的最小特征向量作为数据点的方向矢量,并根据实际测量情况,对基于二次曲面拟合的数据点曲率估算算法进行了改进。对实际测量点云数据,能够较准确地估算出点云方向矢量和曲率,并能形象显示出数据点云的曲率分布。     

Adaptive direct slicing with non-uniform cusp heights for rapid prototyping

Adaptive slicing varies layer thickness by taking the geometry change of the CAD model in the build direction into account to improve surface finish. Direct slicing generates exact slice contours from the original CAD model and avoids an intermediate representation, known as an STL file. At present, most direct slicing approaches are restricted to some CSG solids or some CAD systems. In this paper, an approach toward adaptive direct slicing with non-uniform cusp heights independent of CAD systems for rapid prototyping is presented. First the geometry model is imported into the adaptive direct slicing system from CAD systems using the standard STEP format. Using OpenGL graphics libraries, the solid model is then displayed and the user is prompted to specify the allowable cusp height for each highlighted surface. Lastly, the CAD model is sliced adaptively with different cusp heights (tolerance requirements) for different surfaces. With non-uniform cusp heights, adaptive slicing has a higher efficiency. Implementation details and results are also presented.     

用继承与优化算法精密拼接无序点云

对结构光三维扫描检测系统的拼接技术进行了研究,建立了两个待拼接点云之间的对应关系,提出用继承与优化算法进行点云的精确拼接.阐述了算法原理,通过建模获取拼接过程中的旋转和平移参数,提出并分析J,拼接的实现过程.采用光栅投射式三维扫描仪获取某型号汽车防雨板的6组点云数据,用提出的算法进行点云拼接,利用多分辨率层次精度分析法对拼接结果进行误差分析,并与最临近点迭代法在拼接精度、收敛速度和耗时上进行r比较.实验结果表明:继承与优化算法可实现海量无序点云的精确拼接,拼接的标准偏差<0.10 mm,两点云对拼接时间<2 s,所需迭代次数比ICP方法减少5次以上.     

逆向工程中数字化测量与点云数据处理

逆向工程作为新产品开发和消化、吸收先进技术的重要手段,其应用越来越受到重视。本文主要总了逆向工程中数据测量方法及影响测量精度的因素,从噪声去除、多视对齐、数据光顺、数据精简、数据分割等方面讨论了点云数据的处理方法。