基于倾斜摄影测量的矿产地形三维可视化建模研究

对矿产地形进行三维可视化建模,指导地质勘探,提出基于倾斜摄影测量的矿产地形三维可视化建模方法,采用倾斜摄影测量进行矿产地形的三维图形重建,提取矿产地形的三维边缘轮廓特征量,采用稀疏性的模板特征匹配方法进行矿产地形三维可视化的块匹配,构建矿产地形的边界分布轮廓线,采用不规则三角网重建方法进行矿产地形三维可视化图像的视觉重构,结合倾斜摄影测量方法实现对矿产地形的三维离散点数据重构,提高对矿产地形三维可视化建模的精度。仿真结果表明,采用该方法进行矿产地形三维可视化建模的精度较高,视觉重构效果较好。     

LIDAR点云中多层屋顶轮廓线提取方法研究

在分析LIDAR数据提取建筑物轮廓线现状的基础上,针对多层、非规则屋顶轮廓线提取的难点,提出一种直接基于离散点云的屋顶轮廓线提取方法,该方法主要包括屋顶点的识别,初始轮廓线的提取以及轮廓线的规则化等步骤。最后采用实地数据进行验证,结果表明该方法具有一定的应用前景。     

基于轮廓提取和自适应单元分解的水泥颗粒三维重构

为了得到真实的水泥颗粒初始水化演化模型,提出一种基于轮廓的重构方法和多层次单元分解(MPU)重构方法相结合的水泥颗粒重构方法.首先引入权重函数将提取轮廓的测地线/几何动态轮廓模型和无边缘动态轮廓模型结合到一起,利用动态轮廓模型快速全局最小化方法提取颗粒轮廓,并采用最大类间方差求出边缘阈值,提取出水泥颗粒精确的边缘轮廓线;然后使用平面拟合方法将其转化为带有"法矢"的三维空间点;最后采用改进的MPU方法重构出光顺的三维水泥颗粒.实验结果表明,文中方法比目前水泥材料领域已有的重构方法简单易行,重构效果理想.     

基于多平面重建提取冠脉轮廓线的曲面重建算法

针对三维切割及多平面重建只能获取组织或器官的几何平面信息,无法将弯曲结构的组织或器官展示在单张图片上的问题,实现了基于多平面重建(MPR)提取轮廓线的冠脉曲面重建(CPR)算法.首先,利用多平面重建获取冠脉轮廓的离散点;然后,对离散点进行Cardinal样条插值,获取平滑的轮廓拟合曲线;其次,沿着感兴趣方向对轮廓线进行投影形成扫描曲面;最后,显示扫描曲面对应的心脏体数据,得到冠脉重建曲面.实验结果表明,在绘制速度上,与三维切割法和三维数据场法相比,冠脉轮廓线提取速度提高了每秒4~6帧,绘制时间较短.在绘制质量上,与三维分割法相比,得到的冠脉曲面成像清晰,结构完整,有助于医师对病变的直观分析,能满足实际临床诊疗需求.     

机载激光雷达点云多层次建筑物的三维重建

针对城市区域中常见的多层次建筑物,提出一种基于机载LiDAR点云的三维重建方法。使用优化的随机抽样一致性算法对建筑物LiDAR点云进行面片分割,在面片分割的基础上使用delaunay三角剖分进行轮廓点的检测,对检测出的轮廓点使用新的关键点检测算法提取轮廓线关键点,最后连接关键点并进行规则化处理,完成多层次建筑物的三维重建。实验表明,该方法能有效重建多层次建筑物模型,改进的RANSAC算法能更高效地分割屋顶面片,新的轮廓线关键点提取算法能够较好地提取建筑物的关键点,并能抑制错误轮廓点的干扰,减少伪关键点的生成。     

颌面外科手术三维模型测量和实体模型制造数据提取

文章介绍了一个基于MC算法重构三维头骨模型的颌面外科手术三维模型测量和实体模型制造数据提取系统。该系统使医生可以交互地在三维虚拟模型上提取骨骼特征点,进行比常规方法更精确地测量。还解决了复杂的颌面部插值轮廓线无法提取的问题,可以由三维头骨模型生成符合RP技术要求的轮廓线数据,用于实体模型的制作。     

复杂防雷条件下保护范围及三维模型研究

研究了复杂防雷条件下防雷保护范围模型的重建方法,提出通过保护范围 平面二维轮廓线重建保护范围三维模型的基本思路。通过运用滚球法原理,详细分析了利用 包络原理生成保护范围平面二维轮廓线产生的方法,并给出了相应的数学模型;研究通过二 维轮廓线重建三维模型的方法。由二维轮廓线重建保护范围模型的方案不仅适合于复杂条件 的防雷范围设计,同样也适合简单模型条件下的防雷范围模型,为防雷保护范围设计提供了 一条新的思路。     

基于序列轮廓线的凸面目标全自动重构方法

凸面目标在成像时会表现尺寸较小、纹理稀疏或缺乏、透光、反光等特点.在对其进行三维重构时,传统的立体视觉匹配、主动视觉扫描(激光或结构光)等重构方法无法提供良好的解决方案.针对此类目标的重构难题,本文提出一种基于序列轮廓线的全自动重构方法.该方法首先采集目标的序列轮廓影像,然后提取高精度的轮廓线,再利用多个视角的轮廓线对初始外包体进行空间切割得到目标的初始三维模型,接着根据角度和面积约束对初始三维模型中的面进行合并,最后利用合并后的面重新对外包体进行精细切割得到目标的精确三维刻面模型.通过对凸面目标(注射器针头,直径约为3mm)的重构实验证明,利用本文的方法重构得到的模型角度误差小于0.7,执行时间小于15s,重构过程无需人工干预,能有效解决凸面目标的三维重构及视觉测量难题.     

基于稀疏点云的多平面场景稠密重建

多平面场景是生活中常见的一种场景,然而由于该类场景中常常存在物体表面纹理缺乏和纹理重复的现象,导致从多视图像重建获得的三维点云数据中存在点云过于稀疏甚至孔洞等问题,进而导致以微面片拟合三维点云所得到的重建表面出现平面颠簸现象.针对这些问题,本文提出了一种基于稀疏点云的分段平面场景重建方法.首先,利用分层抽样代替随机抽样,改进了J-Linkage多模型估计算法;然后,利用该方法对稀疏点云进行多平面拟合,来获得场景的多平面模型;最后,将多平面模型和无监督的图像分割相结合,提取并重建场景中的平面区域.场景中的非平面部分用CMVS/PMVS(Clustering views for multi-view stereo/patch-based multi-view stereo)算法重建.多平面模型估计的实验表明,改进的J-Linkage算法提高了模型估计的准确度.三维重建的实验证实,提出的重建方法在有效地克服孔洞和平面颠簸问题的同时,还能重建出完整平面区域.     

一种改进的薄壁文物碎片特征轮廓线提取技术

针对使用一般的边界提取方法提取三维网格模型特征轮廓线不完整问题,提出一 种新的薄壁文物碎片特征轮廓线提取的综合算法。区别了特征轮廓线和轮廓线的概念,引入主 轮廓线和次轮廓线以及二级邻接生长曲面的概念。主轮廓线的提取使用改进的基于边重数判断 的提取方法;提出次轮廓线的一种新的提取方法：首先对三维网格曲面分割并识别断裂面,然 后对断裂面的二级邻接生长曲面进行曲面扫描,提取次轮廓线;最后从主轮廓线和次轮廓线中 得到三维模型的特征轮廓线。使用该算法准确地提取了文物碎片的特征轮廓线,实验结果表明 此方法稳定且准确。     

A random sampling strategy for piecewise planar scene segmentation

We investigate the problem of automatically creating 3D models of man-made environments that we represent as collections of textured planes. A typical approach is to automatically reconstruct a sparse 3D model made of points, and to manually indicate their plane membership, as well as the delineation of the planes: this is the piecewise planar segmentation phase. Texture images are then extracted by merging perspectively corrected input images. We propose an automatic approach to the piecewise planar segmentation phase, that detects the number of planes to approximate the scene surface to some extent, and the parameters of these planes, from a sparse 3D model made of points. Our segmentation method is inspired from the robust estimator ransac. It generates and scores plane hypotheses by random sampling of the 3D points. Our plane scoring function and our plane comparison function, required to prevent detecting the same plane twice, are designed to detect planes with large or small support. The plane scoring function recovers the plane delineation and quantifies the saliency of the plane hypothesis based on approximate photoconsistency. We finally refine all the 3D model parameters, i.e., the planes and the points on these planes, as well as camera pose, by minimizing the reprojection error with respect to the measured image points, using bundle adjustment. The approach is validated on simulated and real data.     

Surface reconstruction with higher-order smoothness

This work proposes a method to reconstruct surfaces with higher-order smoothness from noisy 3D measurements. The reconstructed surface is implicitly represented by the zero-level set of a continuous valued embedding function. The key idea is to find a function whose higher-order derivatives are regularized and whose gradient is best aligned with a vector field defined by the input point set. In contrast to methods based on the first-order variation of the function that are biased toward the constant functions and treat the extraction of the isosurface without aliasing artifacts as an afterthought, we impose a higher-order smoothness directly on the embedding function. After solving a convex optimization problem with a multiscale iterative scheme, a triangulated surface can be extracted using the marching cubes algorithm. We demonstrated the proposed method on several data sets obtained from raw laser-scanners and multiview stereo approaches. Experimental results confirm that our approach allows us to reconstruct smooth surfaces from points in the presence of noise, outliers, large missing parts, and very coarse orientation information.     

Procedure-based component and architecture modeling from a single image

This paper advocates a new component-aware framework to reconstruct 3D architecture from a single image. Different from existing work, our motivation is to obtain a complete set of semantically correct 3D architectural components, which enables part reusability towards rapid model reproduction and facilitates model variation. The core of our system is a novel algorithm to adaptively segment repeated curved stripes (e.g., roof tiles, building floors) into individual elements, based on which 3D dimensions as well as architectural components are derived from a single image. Specially for Chinese architectures, we further devise an interactive method to identify outer columns based on user-specified inner columns. Finally, 3D components are generated and shape rules are derived, from which the buildings and their variants are constructed. Our new component-aware framework minimizes the use of data resource (i.e., one single image) and emphasizes component utility during rapid 3D architecture reproduction by advocating a component-aware approach.     

Fast construction of plant architectural models based on substructure decomposition

Plant structure,representing the physical link among different organs,includes many similar substructures.In this paper,a new method is presented to construct plant architectural models of most plant species.The plant structure is decomposed into a stem,a set of lateral substructures and a terminal substructure,which is called substructure decomposition;then based on substructure decomposition,the plant structures are expressed in an iterative way;and further the derivative formula is employed to compute the number of organs in plant structures to get the geometrical sizes of 3D plant organs by borrowing Hydraulic Model.Using 3D organs,a substructure library is built.Based on the substructures stored in substructure library,one can construct 3D plant structure according to certain topological and geometrical rules.The experiments with different plant species are included in this paper to demonstrate the validity of the new method for constructing plant structures.The experimental results show that the approach follows botanical knowledge with high efficiency in constructing plant structures of most plant species.In addition,this method enables users to check the detail information of plant structure.     

3D reconstruction of buildings from LiDAR data considering various types of roof structures

In this article, a different approach has been proposed to detect and reconstruct buildings in 3D space. In this regard, some potentially primary features were produced and a genetic algorithm (GA) was employed in order to select the optimum features. The selected features were utilized to detect the building by using k-nearest neighbour (k-NN) algorithm. The detection results were used as inputs of reconstruction procedure. The proposed approach for 3D reconstruction consists of three main steps: roof planes were separated in the first step. Then, the corners of each plane boundary were extracted in order to provide the roof reconstruction possibility. Finally, the walls were reconstructed and merged to roofs and the final 3D model was obtained. Results evaluation indicated that the average value of buildings detection in the test areas was 87.84% in Quality. Moreover, the average value of buildings reconstruction in the test areas was 76.95% in object based Quality, and 95.66% in Quality of building planes that were greater than 25 m², respectively. Also, the average of altimetric and planimetric RMS of the test areas were 0.3 m and 0.75 m, respectively.     

A method of automatic extraction of parameters of multi-LoD BIM models for typical components in wooden architectural-heritage structures

A building-information-modeling (BIM) model, which is established based on high-fidelity point-cloud data, can be well used to preserve architectural heritage. Two related critical issues for this conservation are multiple-level-of-detail (multi-LoD) parametric models that emphasize different protection requirements for typical components, and a method for automatically extracting the corresponding parameters from a high-fidelity point cloud. Taking typical Chinese wooden architectural-heritage structures as an example, multi-LoD principles for typical components without damage are proposed according to the different requirements. Then, a framework of multi-LoD parametric models was developed and implemented in BIM. Based on this, a method for automatically extracting the abovementioned parameters is developed and the critical parameters of this method are recommended. To validate the reliability and efficiency of this method, the parameters of multi-LoD models of typical components are extracted. The results indicate that the relative and absolute errors of values of such parameters are mostly less than 2% and 0.5 mm, respectively. Moreover, this method is capable of extracting parameters from millions of point-cloud data within 7 min, thus validating the high efficiency and reliability of the proposed method.     

Live 3D Video in Soccer Stadium

This paper proposes a method to realize a 3D video system that can capture video data from multiple cameras, reconstruct 3D models, transmit 3D video streams via the network, and display them on remote PCs. All processes are done in real time. We represent a player with a simplified 3D model consisting of a single plane and a live video texture extracted from multiple cameras. This 3D model is simple enough to be transmitted via a network. A prototype system has been developed and tested at actual soccer stadiums. A 3D video of a typical soccer scene, which includes more than a dozen players, was processed at video rate and transmitted to remote PCs through the internet at 15–24 frames per second.     

基于图形理解的建筑结构三维重建技术

在实际应用中,现有的三维重建算法无法有效地应用到建筑领域.以建筑行业的基本制图规范为指导,结合图形识别理解和人工智能技术,提出了一种基于图形理解的建筑结构三视图自动识别与重建方法.该方法以自动识别建筑构件的轮廓为基础,通过对符号、注释等语义信息的理解与综合以及对相邻实体间拓扑关系的分析,快捷、有效地完成对建筑物的整体重建.实验结果表明,在少量人机交互的基础上,该方法对实际的建筑图是有效的.     

基于测地回归模型的颅面复原

颅面复原是指根据一个未知颅骨的特征预测出对应的面貌,在考古研究、医学整容、刑事案件调查等领域有重要应用.为解决颅面复原过程中存在着数据量大、需要大量标定特征点的手工工作及颅面特征点定义困难的问题,针对三角网格表示的三维颅骨和面皮模型,将颅面模型用从鼻尖出发的一组测地线表示,提出了基于测地回归的颅面复原方法.该方法首先从...