三维

3D reconstruction is one of the most challenging problems in computer vision. This project uses CAD model to increase the accuracy,robustness and speed of 3D reconstruction and five different works are conducted: 1 research on space curve 3D reconstruction method based on high curvature point nad turning point; 2 free-form curve 3D reconstruction based on B-spline and BURBS; 3 Active surface affine reconstruction based on single frame; 4 Matching and reconstruction of 3D sparse points; Data fitting method based on the iteration of nearest point.

三维重建是计算机视觉中公认的难题，本项目使用CAD模型来提高三维重建的精度、稳健性和速度的思路，围绕着空间曲线和曲面的三维重建和匹配开展了以下五个方面的研究：1）基于高曲率点和拐点分割的空间曲线三维重建方法研究；2）基于B样条和NURBS表示的空间自由曲线三维重建方法研究；3）基于单帧图像的有源曲面仿射重建；4）三维离散数据的匹配与重建；5）基于迭代最近点的数据拟合方法。

The coordinate transformation and numerical integration was executed on the discretized tetrahedral elements based on which the 3-D MT vector-finite element method was implemented. A whole computation framework for 3-D vector-finite element method with unstructured mesh was given. Based on this some typical models were tested which has demonstrated that our algorithm could distinctively avoid problems caused by the fake solution and both the accuracy and efficiency were enhanced which made our algorithm has a bright future for further application.3. According to theory of Sobolev vector space and the discretization of Helmholtz space, the error estimate which was suitable for 3-D MT vector-finite element modeling was deduced by which the procedure of adaptive technique was guaranteed.4. Based on the fully unstructured tetrahedralization and optical strategy, the 3-D magnetotelluric h-adaptive vector-finite element method was presented through combining the error estimate. With this work, the accuracy and creditableness for 3-D MT complicatedly modeling was guaranteed.5. The 3-D magnetotelluric h-adaptive vector-finite element algorithm with unstructured mesh was implemented.

针对非结构化的四面体单元，采用坐标变换和数值积分方法，实现了MT三维矢量有限单元分析，建立起基于非结构化网格的三维MT矢量有限元计算流程，并对典型模型和国际标准电磁模型进行了数值模拟，结果对比和分析表明，基于非结构化网格的三维MT矢量有限元不仅消除了节点型有限元的伪解，而且具有很高的计算精度和速度，有广阔的应用前景。3、根据Sobolev函数的向量空间和Hemlholtz空间的分解，推导出基于残差的三维大地电磁矢量有限元后验误差估计公式，为三维大地电磁自适应矢量有限元数值模拟的实现奠定了基础。4、在完全非结构化四面体单元剖分及优化基础上，结合三维大地电磁矢量有限元后验误差估计公式，提出了基于非结构化网格的三维大地电磁h-型自适应矢量有限元计算策略，保证了对复杂大地电磁模型数值计算的精度和可靠性。5、实现了基于非结构化网格的三维大地电磁h-型自适应矢量有限元计算流程，对典型模型和国际标准电磁模型进行了数值模拟。

In view of the demand of computerized eye-glasses-wearing system, an approach to match two dimensional (2D) human head images with 3D glasses frame graphics in 3D space has been proposed. First, the operator points out three or six key points, which are used to locate 3D eye-glasses frame, on the human face image displayed on the computer screen interactively.

论文针对计算机化眼镜配戴系统的要求，提出了一种在三维空间匹配二维人头图象和三维眼镜图形的方法，即通过人机交互方法，由操作者在显示于计算机屏幕上的二维人头图象上，指出确定三维眼镜位置的三个或六个关键点的二维坐标及其三维对应点，从而可以近似求出人面朝向及视点坐标，再用相应位置上的三维通用人头图形替换二维人面图象，计算人面和眼镜间的遮挡关系并显示，从而完成二者的空间匹配。

Then loading the models in both feet standing body posture, gain stress distribution figures of the models, record stress numerus of weight bearing dome, anterior wall and posterior wall in condition of different areas posterior wall fractures. Results 1、Normal 3D finite element model of acetabulum Obtain 3D visible model of pelvic by restituting CT fault images with Mimics10.01; gain 3D visible model of haunch bone by splitting the pelvic model with FreeForm software.

结果 1、正常髋臼三维有限元模型通过Mimics10.01软件对骨盆断层CT图像进行三维重建，获得骨盆三维可视模型；然后在FreeForm软件系统中对模型进行修饰，平滑模型表面，设定各受力面与界线；并成功截取出髋骨三维可视模型；再将模型导入Ansys8.0中进行各部分材料赋值及网格划分，建立了正常髋臼三维有限元模型，共含节点69440个，三维四面体单元45209个。

3D solid visualization class and 3D surface visualization class have thefunction of visualizing the scalar and vector. Moreover, 3D surface visualization classcan produce 2D data file. Picture class provides function of marking on the picture.

其次，运用 VTK 标量和矢量可视化技术实现了三维数据场的矢量和标量可视化，包括三维体数据场标量和矢量单视区显示及三维体数据场剖面标量和矢量单视区/四视区显示；使三维面数据可视化类具备了三维面数据场标量单视区显示功能和剖切面生成二维数据文件的功能，并且在三维面数据可视化类中内嵌了数据预处理算法。

The key technical problem of roughcast in the virtual machining environment is to handle three-dimensional shape, which includes the method of setting up geometrical model of three-dimensional shape of roughcast , the data structure that displays three-dimensional shape, process compiling three-dimensional shape, etc.

在虚拟加工环境中，毛坯的三维图形处理主要包括毛坯三维图形的几何建模方法、三维几何图形的显示数据结构和三维图形的编译处理等，提出了一种能够满足虚拟加工环境的三维图形的数据结构，给出了实现方案，实现了三维图形的环境调入和毛坯的对接等功

The invention discloses the anti-fake method by using the 3D stereoscopic icon, which can show 3D stereoscopic icons and hide the confidential image information in the 3D stereoscopic icons, comprising the following steps of: obtaining parameters, inputting confidential images and grey images; then using a visual divided storage method to build n visual divided storage images, using a 3D stereograph method to build n 3D stereoscopic images, adjusting the sizes of the n visual divided storage images and the n 3D stereoscopic images and dividing the n visual divided storage images and the n 3D stereoscopic images into a plurality of basic blocks; finally random choosing the basic blocks in the corresponding positions in the visual divided storage images or the 3D stereoscopic images to build n 3D stereoscopic and visual divided storage icons, etc.

为了既能三维立体显示图标，又能在三维显示的图标中隐藏秘密图像信息，本发明公开了利用三维立体图标防伪的方法，包括获取参数，输入秘密图像和灰度图像；然后利用可视分存方法构造n幅可视分存图，利用三维立体画方法构造n幅三维立体图，调整所述n幅可视分存图和所述n幅三维立体图大小并分成多个基本块；最后随机选取可视分存图或三维立体图中相应位置处的基本块，构造n幅三维立体可视分存图标等步骤。

By the numerical analysis of the ultimate bearing capacity of two- and three-dimensional footing under combined loading mode, the results show that the effect of alteration of underground water lever on bearing capacity of three-dimensional foundation is more than on that of two-dimensional foundation, the effect of soil gravity on bearing capacity of three-dimensional foundation is more than that of two-dimensional foundation. At the same time, the numerical calculation process of ABAQUS indicates that whether the geostatic stress is balanced or not has great effect on the numerical convergence. In view of the effect extent of dilation angle of soil on the ultimate bearing capacity and the failure pattern of foundation, the condition of two-dimensional plane strain is higher than that of the three-dimensional foundation.

复合加载模式下，通过对三维矩形基础和二维条形基础进行数值分析，结果表明：黏性土地基中地下水位的变化，对三维矩形基础极限承载力的影响程度高于对二维条形基础的影响；在求解地基极限承载力的过程中，土体重度对三维矩形基础的影响明显高于对二维条形基础的影响，同时ABAQUS数值计算过程表明，地基土体地应力能否合理平衡，对数值求解过程中是否收敛有很大的影响；就土体剪胀角对地基极限承载力的影响程度而言，二维平面应变情况下土体剪胀角对极限承载力的影响比三维情况下的影响更明显，即土体剪胀角对条形基础极限承载力的影响高于三维矩形基础。

The model is used to count several engineering example.Comparing the 2D numerical results with electrical simulations,it is proved that the seepage model can count the seepage surface and the location of search results in dam well and truly.Comparing 3D results with 2D results at close section,the results are tally well except the difference of meshes and the interceptive position of section.So the model is credible and has upper practicality value.In the projects,the spending of account is increasing because of the bigger account field.

将饱和—非饱和渗流模型用于几个工程实例，其二维模拟结果同渗流电模拟实验结果进行了对比，证明本文的渗流模型完全能够准确计算坝内的渗流浸润线以及逸出面位置；三维数值模拟的结果在相近剖面处与二维结果作了比较，考虑三维网格剖分及剖面截取位置的差异，三维结果与二维结果吻合的较好，证明本文的饱和—非饱和渗流模型在三维计算中是可靠的且具有较高的实用价值。

The research of making use of 2D image to form the 3D objects has been a field paid attention to very much all the time. In this work, we propose the synthetic way of a 3D human head, based on image information of multiple-view, cooperate with a Generic 3D head model. By choosing the feature points from images to estimation of position of the camera, and then dealing with information reconstruction of human head and feature point relationship between images and 3D object. Subsequently, applying interpolation technique, the head model is deformed from original model to fit the computed feature points by using Radial Basis Function. Make the form of the head model of people after deformation in conformity with true face of people. Finally, execute 3D head texture mapping and reach the function of translation and rotation with arbitrary direction.

利用二维影像来合成三维物件的研究一直是很受到重视的一个领域，在本论文中，我们提出一个三维立体人头的合成方式，以多张不同视角的影像资讯为基础，配合一个事先准备好的通用三维人头模型(generic 3D head model)，透过二维影像中特徵点的选取来估算出拍摄的角度、方位，进而处理人脸模型五官的资讯重塑与特徵点对应的关系，接著利用内差技术，使用放射基底函数对三维通用人头模型作形变，使形变过后人头模型的形状与真实人脸相符，最后对三维通用人头模型作材质贴图，并可对此模型作任意角度、方向的旋转平移，以期达到类似照镜的功能。

I'm gonna go pack my regular, long bathing suit.

我要去打包我那不起眼的超长泳裤

Random number generators are used in many areas, with the improvement of the computational ability of the computer.

随着计算机计算能力的不断提高，随机数发生器在许多领域有了更加广泛的应用。