一种基于Multi-Egocentric视频运动轨迹重建的多目标跟踪算法

Egocentric视频具有目标运动剧烈、遮挡频繁、目标尺度差异明显及视角时变性强的特点，给目标跟踪任务造成了极大的困难。本文从重建不同视角Egocentric视频中各目标的运动轨迹出发，提出一种基于Multi-Egocentric视频运动轨迹重建的多目标跟踪算法，该方法基于多视角同步帧之间的单应性约束解决目标遮挡和丢失问题，然后根据多视角目标空间位置约束关系通过轨迹重建进一步优化目标定位，并采用卡尔曼滤波构建目标运动模型优化目标运动轨迹，在BJMOT、EPLF-campus4数据集上的对比实验验证了本文算法在解决Multi-Egocentric视频多目标跟踪轨迹不连续问题的有效性。     

动态识别三维几何约束冲突的方法研究

基于装配几何特征的广义几何约束图,避免了传统几何约束图的超图性质和模糊性,为几何约束满足问题提供了一个清晰的分析模型。文中以此模型来分析产生约束冲突的原因。空间分析法定义和推导了约束满足空间约束满足条件,提出了自由空间和自由度的计算方法,并据此在动态满足三维几何约束的过程中识别约束冲突,明确指出产生约束冲突的原因。     

结合三维几何形状信息和二维纹理的3D目标匹配

为了实现场景中三维目标与模型之间的匹配,提出了一种结合三维几何形状信息和二维纹理的三维目标匹配方法。首先提取场景中深度图像的尺度不变特征变换(SIFT)特征,用SIFT算法与三维模型重建时所用到的一系列2.5维深度图像进行一一匹配,找到与场景中目标姿态最为相似的深度图像,提取此深度图像的三维几何形状特征与模型进行匹配,实现模型的初始化,即将模型重置到与场景目标相接近的姿态。最后用融合二维纹理信息的迭代就近点（ICP）算法实现场景中目标与模型之间的匹配,从而得到场景中三维目标的准确姿态。实验结果验证了方法的可行性与精确性。     

一种基于双幅图像的物体3维重建

随着虚拟现实技术的发展,对物体进行快速有效建模逐渐成为了研究热点。基于图像建模技术是一种有效的建模方法,其最大的优点是可以利用图像中的信息直接建立物体的几何模型,并且可以迅速构建出具有“照片级”真实感的3维模型。常规的方法是先求解摄像机内外参数,再求解空间中各点在世界坐标系下的位置。提出了一种基于两幅图像利用灭点性质进行建模的新方法。该方法首先利用灭点属性推导出了相机中心在世界坐标系下的坐标位置的解析表达式,再利用两幅图像之间的空间几何约束关系,从空间约束关系中反算出物体各点的空间位置从而重建出物体几何模型,最后进行纹理映射。在物体重建的过程中只利用到了相机内参数,而跳过了对相机外参数的标定,实验结果表明,该方法简单可靠,可用于实现对拥有规则表面的物体3维重建。     

基于Kinect与网格几何变形的人脸表情动画

面部表情重建的实时性与重建效果的真实性是人脸表情动画的关键问题,提出一种基于Kinect人脸追踪和几何变形技术的面部表情快速重建新方法。使用微软的Kinect设备识别出表演者面部并记录其特征点数据,并利用捕捉到的特征点建立覆盖人脸的网格模型,从中选取变形使用的控制点数据,由于Kinect可以实时地自动追踪表演者面部,由此实现了利用三种不同变形算法对目标模型实时快速重建。实验结果表明,该方法简单易实施,不用在表演者面部做任何标定,可以自动地将人脸表情动作迁移到目标模型上,实现人脸表情快速重建,并且保证目标模型表情真实自然。     

基于关联约束的工程图纸表信息识别与重组

产品技术信息是实现产品数据集中管理的基础，在工程图中通常以表的形式存在．根据对表的内部约束和外部约束的分析，建立表的几何关联模型与逻辑关联模型；在此基础上，通过一系列算法实现了表信息的识别、重组，并依照表与数据库之间的映射关系，将表信息直接录入数据库，同时重建信息之间的关联，为产品设计制造提供了数据基础。     

面向软件产品线中特征选择的多目标优化算法

软件产品线中产品定制的核心是选择合适的特征集.由于多个非功能需求间往往相互制约甚至冲突,特征选择的本质是多目标优化过程.优化过程的搜索空间被特征间错综复杂的依赖和约束关系以及明确的功能需求大大限制.另外,有些非功能需求有明确的数值约束,而有些则仅要求尽可能得到优化.多样的非功能需求约束类型也给优化选择过程带来极大挑战.本文提出一种含修正算子的多目标优化算法MOOFs.文中首先设计特征间依赖和约束关系描述语言DCF-DL来统一规范特征选择过程中必须遵守的规则.所有的非功能需求都转化为优化目标,相关的数值约束则作为优化过程中特征选择方案的过滤器.另外,设计了修正算子用于保证选择出的特征配置方案必满足产品线的特征规则约束.通过与四个常用的多目标优化算法在四个不同规模的特征模型上的运行结果做对比,证明本文的方法能够更快地产生满足约束的优化解,且优化解具备更好的收敛性与多样性.     

基于特征CAD中的特征定位关系及算法

提出了以实体模型、测量要素、定形尺寸、定位尺寸、行为约束为构造成份的特征模型；为了捕获ＣＡＤ中特征之间的相互定位关系，建立了特征父子关系算子结构；提出了特征父子关系图的概念用于表达零件模型；给出了访问父子关系图的算法，讨论了算法实现的例子。     

双目立体视觉的三维人脸重建方法

创建逼真的三维人脸模型始终是一个极具挑战性的课题.随着三维人脸模型在虚拟现实、视频监控、三维动画、人脸识别等领域的广泛应用,三维人脸重建成为计算机图像学和计算机视觉领域的一个研究热点.针对这一问题,提出一种基于双目立体视觉的三维人脸重建方法,重建过程中无需三维激光扫描仪和通用人脸模型.首先利用标定的2台摄像机获取人脸正面图像对,通过图像校正使图像对的极线对齐并且补偿摄像机镜头的畸变;在立体匹配方面,选择具有准确可靠视差的人脸边缘特征点作为种子像素,以种子像素的视差作为区域生长的视差,在外极线约束、单调性约束以及对应匹配的边缘特征点的约束下,进行水平扫描线上的区域生长,从而得到整个人脸区域的视差图,提高了对应点匹配的速度和准确度;最后,根据摄像机标定结果和立体匹配生成的视差图计算人脸空间散乱点的三维坐标,对人脸的三维点云进行三角剖分、网格细分和光顺处理.实验结果表明,该方法能够生成光滑、逼真的三维人脸模型,证明了该算法的有效性.     

多视觉环境下联合多约束的三维目标体重建

提出了一种联合多种约束条件的体元重建方法。首先将场景空间离散为体元,并从多视获取的二维图像中提取目标轮廓信息;其次利用侧影轮廓约束及颜色一致性约束,恢复目标三维信息;针对重建模型存在漂浮体元和毛刺现象,提出加入平滑约束解决该问题;最后提出模型修复算法对模型表面空洞进行填补,解决重建模型空洞问题。实验结果表明,该方法可重建颜色纹理信息准确的目标外形,且模型光滑细腻,表面空洞得到有效填补。     

一种适用于特征造型的参数化设计方法

本文提出了一种面向特征造型的参数化设计方法，该方法对三维几何约束在初始设计阶段采用高层表示，并基于面向特征造型的高层几何约束模型与约束传播实现尺寸驱动几何，从而能够有效地支持特征设计，初始设计。     

A set of standard modeling commands for the history-based parametric approach

The current version of STEP standard cannot exchange the parametric information of CAD models. Only pure boundary representations that cannot be parametrically edited are transferable [Geometric Modeling: Theory and Practice (1997)]. There are two approaches for the exchange of design intents such as parameters, features, and constraints. The first is an explicit approach based on constraints between pre-defined parameters and features. The second is a procedural approach based on the sequence of operations issued to construct the models. The authors have previously proposed a macro-parametric approach [International Journal of CAD/CAM 2 (2002) 23], which is a variation of the procedural approach. In this approach, CAD models can be exchanged in the form of macro files, which include the history of modeling commands. To exchange CAD models using the macro-parametric approach, a set of standard modeling commands should be defined. This paper introduces a set of standard commands and explains the process of developing the set.     

Introducing the sequential linear programming level-set method for topology optimization

This paper introduces an approach to level-set topology optimization that can handle multiple constraints and simultaneously optimize non-level-set design variables. The key features of the new method are discretized boundary integrals to estimate function changes and the formulation of an optimization sub-problem to attain the velocity function. The sub-problem is solved using sequential linear programming (SLP) and the new method is called the SLP level-set method. The new approach is developed in the context of the Hamilton-Jacobi type level-set method, where shape derivatives are employed to optimize a structure represented by an implicit level-set function. This approach is sometimes referred to as the conventional level-set method. The SLP level-set method is demonstrated via a range of problems that include volume, compliance, eigenvalue and displacement constraints and simultaneous optimization of non-level-set design variables.     

基于深度强化学习的微电网在线优化调度

提出一种基于深度强化学习的微电网在线优化调度策略.针对可再生能源的随机性及复杂的潮流约束对微电网经济安全运行带来的挑战,以成本最小为目标,考虑微电网运行状态及调度动作的约束,将微电网在线调度问题建模为一个约束马尔可夫决策过程.为避免求解复杂的非线性潮流优化、降低对高精度预测信息及系统模型的依赖,设计一个卷积神经网络结构学习最优的调度策略.所提出的神经网络结构可以从微电网原始观测数据中提取高质量的特征,并基于提取到的特征直接产生调度决策.为了确保该神经网络产生的调度决策能够满足复杂的网络潮流约束,结合拉格朗日乘子法与soft actor-critic,提出一种新的深度强化学习算法来训练该神经网络.最后,为验证所提出方法的有效性,利用真实的电力系统数据进行仿真.仿真结果表明,所提出的在线优化调度方法可以有效地从数据中学习到满足潮流约束且具有成本效益的调度策略,降低随机性对微电网运行的影响.     

Mobile robot relocation from echolocation constraints

This paper presents a method for relocation of a mobile robot using sonar data. The process of determining the pose of a mobile robot with respect to a global reference frame in situations where no a priori estimate of the robot's location is available is cast as a problem of searching for correspondences between measurements and an a priori map of the environment. A physically-based sonar sensor model is used to characterize the geometric constraints provided by echolocation measurements of different types of objects. Individual range returns are used as data features in a constraint-based search to determine the robot's position. A hypothesize and test technique is employed in which positions of the robot are calculated from all possible combinations of two range returns that satisfy the measurement model. The algorithm determines the positions which provide the best match between the range returns and the environment model. The performance of the approach is demonstrated using data from both a single scanning Polaroid sonar and from a ring of Polaroid sonar sensors     

An approach to developing complex database schemas using form types

In this paper we consider an approach to developing complex database schemas. Apart from the theoretical model of the approach, we also developed a CASE tool named Integrated Information Systems*Case, R.6.2 (IIS*Case) that supports the practical application of the approach. In this paper the basis of our approach to the design and integration of database schemas and ways of using IIS*Case is outlined. The main features of a new version of IIS*Case, developed in Java, are described. IIS*Case is based on the concept of ‘form type’ and supports the conceptual modelling of a database schema, generating subschemas and integrating them into a relational database schema in 3NF. IIS*Case provides an intelligent support for complex and highly formalized design and programming tasks. Having an advanced knowledge of information systems and database design is not a compulsory prerequisite for using IIS*Case. IIS*Case is based on a methodology of gradual integration of independently designed subschemas into a database schema. The process of independent subschema design may lead to collisions in expressing real‐world constraints. IIS*Case uses specialized algorithms for checking the consistency of constraints embedded in a database schema and its subschemas. This paper briefly outlines the application of the process of detecting collisions, and actions the designer may take to resolve them. Copyright © 2007 John Wiley & Sons, Ltd.     

On the Use of a Mixed Integer Non-linear Programming Model for Refrigerant Design

In this paper, a novel mixed integer non-linear programming model for single component refrigerant design is presented. At the heart of the approach is a new formulation for structural feasibility that allows multiple bonds, connectivity and isomers. The strategy defines a set of structural groups (consisting of atoms), subsets of which are combined to form refrigerant molecules. Molecules formed this way must obey structural and stability constraints. The design objective is to build a refrigerant molecule that has desired physical properties and performance characteristics. These attributes are formulated as mathematical programming constraints and performance objectives which involve both continuous and integer variables. With the current renewed interest in the environment, the suggested approach is applied to refrigerant design with an environmental constraint. The results indicate the viability and the flexibility of the approach.     

基于仿真和知识的广义动态约束网络技术

针对并行协同设计中的参数不确定性,将普通的约束网络扩展为广义动态约束网络,以对设计中的不确定性信息进行管理.建立了包含领域级约束和知识级约束的广义动态约束网络模型;提出了基于仿真分析和自适应响应面法的领域级约束建模的有效方法,并提出模糊-粗糙集算法,对仿真结果进行数据挖掘,实现了知识级约束获取;基于模板技术给出了广义动态约束网络中各种约束的统一表示方法;构造了有效的约束冲突求解策略和一致性求解算法,求出一致性设计区间.最后通过设计实例验证了文中方法的有效性.     

Adding inclusion dependencies to an object-oriented data model with uniqueness constraints

We study an object-oriented data model that allows to express both uniqueness constraints and inclusion dependencies as semantic constraints. The data model is based on a subset of F-logic. Uniqueness constraints comprise path functional dependencies which generalise functional dependencies and reflect the navigational power of object-oriented query languages. As inclusion dependencies, we consider explicit class inclusion constraints, besides inclusions required by class hierarchies, and onto constraints that enforce reachability of objects. For these classes of semantic constraints we present an axiomatisation and prove its inference rules to be correct and complete with respect to general logical implication, leaving the decision problem open. The completeness proof combines the known construction for path functional dependencies alone with a possibly infinite model generation process to enforce onto constraints. The results prepare the grounds for normal forms in object-oriented data models and subsequently for computer aided object-oriented database design, following the decomposition approach for the relational data model. Beyond the application for schema design, the achievements could also be exploited for related tasks like semantic query optimisation and mediated data integration within a variety of graph based data models. Received: 11 October 2000 / 27 January 2003     

Ontology-based feature mapping and verification between CAD systems

Data interoperability between computer-aided design (CAD) systems remains a major obstacle in the information integration and exchange in a collaborative engineering environment. The use of CAD data exchange standards causes the loss of design intent such as construction history, features, parameters, and constraints, whereas existing research on feature-based data exchange only focuses on class-level feature definitions and does not support instance-level verification, which causes ambiguity in data exchange. In this paper, a hybrid ontology approach is proposed to allow for the full exchange of both feature definition semantics and geometric construction data. A shared base ontology is used to convey the most fundamental elements of CAD systems for geometry and topology, which is to both maximize flexibility and minimize information loss. A three-branch hybrid CAD feature model that includes feature operation information at the boundary representation level is constructed. Instance-level feature information in the form of the base ontology is then translated to local ontologies of individual CAD systems during the rule-based mapping and verification process. A combination of the Ontology Web Language (OWL) and Semantic Web Rule Language (SWRL) is used to represent feature classes and properties and automatically classify them by a reasoner in the target system, which requires no knowledge about the source system.