增强现实环境下的CPS虚拟装配系统研究与实现

基于双目视觉原理，利用Unity3D游戏引擎开发了一种增强现实环境下的信息物理融合(cyber-physical systems, CPS)虚拟装配系统。通过研究快速定位算法识别特征标记点建立虚拟模型与实物装配单元的配准关系，融合虚拟模型投影图像与实物装配体图像实现虚拟装配仿真，实时输出在视频眼镜结构中显示。经过试验验证，该系统满足CPS实时性、稳定性以及信息世界与物理世界融合的要求。     

基于多目视觉技术的多点触摸屏系统的实现方法

设计了一种基于多目视觉技术的多点触摸屏系统。系统使用三个摄像机分别置于显示屏的左右两端和前方,利用计算机图像处理技术使普通显示屏具备触摸功能。为了有效减少图像处理时间,提高指尖坐标精度,提出了基于卡尔曼滤波器的指尖跟踪算法,通过卡尔曼滤波器预测指尖位置并以预测指尖点为中心划出矩形ROI(Region of Interest,感兴趣区域)进行指尖检测算法处理。实验证明,该系统能实时运行在嵌入式平台,并可实现较高的点击成功率,对于实现基于计算机视觉的触摸屏技术有着重要的研究意义。     

单目视觉与倾角仪组合优化的位姿测量系统

为了提高单目视觉实时测量双护盾隧道掘进机前后盾相对位姿的精度,引入高精度倾角传感器与单目视觉构成一种组合测量系统。该系统将两个倾角传感器分别与视觉传感器和特征点系统固定连接,通过倾角传感器提供的多个角度约束,结合单目视觉实现掘进机前盾体相对于后盾体位姿的更高精度测量。仿真实验表明该系统是可行的,并且具有理想的精度。搭建了模拟双护盾隧道掘进机位姿变化的实验平台,利用全站仪进行精度验证。结果表明系统的测量精度优于3mm,相对于单目视觉测量方法来说,测量精度有了显著提升,可以满足隧道施工中双护盾隧道掘进机位姿的精密测量需求。     

融合单目视觉的激光扫描投影系统自标定方法

为解决现有激光扫描投影系统中需要手动标定的问题,提出了融合单目视觉的激光扫描投影系统自标定方法。首先建立了融合单目视觉技术的激光扫描投影系统模型,给出了融合单目视觉的激光扫描投影系统自标定方法;其次建立了单目相机与激光扫描投影系统、被投影物体之间的数学模型;最后运用Matlab软件进行模拟仿真,证明融合单目视觉的激光扫描投影系统自标定方法可以解决上述问题。同时为验证融合单目视觉的激光扫描投影系统自标定方法的精度是否满足激光扫描投影系统的需求,将Matlab软件求解的实验结果与SA软件求解的实验结果进行对比,并分析精度与距离、角度等外部因素的关系。仿真实验结果表明:当距离一定时,单目相机旋转角度为55°时,单目相机距离投影仪和被投影物体的范围为1.5~3 m,且满足投影仪的精度要求。同时通过自主搭建新型激光扫描投影系统,证明了融合单目视觉的激光扫描投影系统自标定方法具有实际可操作性和可行性。     

基于单目视觉和圆结构光的目标姿态测量方法

为解决实际工程中无法在被测目标表面设置固定特征来配合单目视觉系统实现目标姿态测量的难题,将圆结构光源引入单目视觉系统中。通过建立圆结构光的视觉姿态测量模型,提出了一种基于单目视觉和圆结构光的目标姿态测量方法。利用图像处理获取不同姿态下目标表面的结构光光条图像的数学参数;然后将其输入到姿态测量算法中,得到目标表面结构光光条的法向量;最后利用目标表面的参考点与结构光光条中心之间的距离约束关系确定唯一解。实验结果表明,该方法有效可行,测量误差平均在0.5°以内,更便于机器人抓取等工程领域的应用。     

光电成像探测中目标方位测量方法

为了在光电成像探测目标的同时给出目标相对探测器的方位信息,探讨了一种基于单目视觉的运动目标方位测量方法。根据投射成像原理,推算出二维目标像点坐标与目标的空间三维位置之间的映射关系,建立了单目视觉测量目标方位的数学模型。结合帧间差分法和KLT方法检测静止背景下和变化背景下的运动目标,并对目标特征点进行亚像素定位。试验和仿真结果分析表明,该方法能够有效提取目标的特征点,可对目标较为准确地定位,并通过单目视觉测量出目标方位信息,误差控制在8%的范围内。     

基于单目视觉的矩形靶面弹着点测量

目前机器视觉技术广泛应用于弹着点坐标的测量领域,针对双目视觉技术标定、演算复杂的问题,提出一种基于单目视觉技术的弹着点测量方法。通过图像处理技术定位出靶面4个角点以及弹着点像素坐标,采用矩形P4P方法解算出靶面坐标相对于相机坐标系的位姿,根据单目视觉成像原理计算出弹着点在靶面的实际坐标。根据该方法成功测量出了实验中靶面弹孔的坐标,最大测量误差为2.3 mm。结果表明,该方法可以快速、精确测量靶面与相机的位姿关系以及弹着点的坐标。     

基于多向联合交汇的柔性视觉形貌测量新方法

针对大尺寸物体形貌测量应用,设计一种视觉形貌测量新方法。该方法是将两套没有公共视场的单目视觉传感器组合起来联合交汇,分别用于位姿测量和单目多位置交汇测量。位姿测量通过求解PnP问题为单目多位置交汇测量提供位姿信息。单目多位置交汇测量利用位姿测量提供的辅助信息通过光束平差解算出待测物的形貌。分析了系统原理,给出位姿测量PnP算法实现过程,介绍了非公共视场相机的标定方法原理,分析了单目多位置交汇测量的原理并给出实现方法。最后进行了测量方法的验证实验,实验结果表明,在3 m×3 m的工作范围内系统测量精度为5 mm,证明该方法可以用于大尺寸物体形貌测量,且结构简单轻便,具有很强的灵活性。     

基于单目视觉的曲面弧度测量方法

针对异形件曲面弧度的测量问题,提出了一种基于单目视觉的曲面弧度测量方法.首先,用基于三色LED环形光源和CCD彩色相机组成的单目视觉系统采集单张包含被测曲面表面特性的彩色图像;其次,分析该图像的彩色分布特征、灰度变化规律和相邻区域特征的相关性,设计分割方法,完成目标曲面的分割;然后,基于单目视觉系统成像模型,结合材质与实验环境相关参量,推导出被测表面高度计算方程;最后,解出相邻像素点间的高度累加分量,进而还原被测点的高度,在此基础上,建立曲面弧度测量模型并完成曲面弧度测量.分别对3D打印曲面样品、螺钉螺纹、SIM卡槽进行了曲面弧度测量实验,与激光三角法测量结果对比,所提方法误差小于±3.6%,测量准确度为0.000 1rad,曲面弧度测量时长小于0.6s.实验结果表明,该方法对测量物体边缘、高度突变,高度渐变等类型的曲面都具有良好的适应性,在实现较高准确度测量的同时,具有较快的测量速度.     

视觉技术辅助的无人机自主着陆组合导航研究

为提高无人机自主着陆过程中导航系统的自主性与精确性,设计了一种视觉辅助惯导组合导航方法。该方法以惯导误差方程为过程方程,以着陆过程中单目摄像机2个时刻所得地面特征点投影之间的双视图几何约束为量测方程,构建了非线性滤波器;利用SR-UKF方法实现了惯导误差估计,提高算法效率的同时有效地避免了UKF中由于矩阵开方运算导致的滤波失效;最后根据估计结果校正了惯导导航数据。仿真结果表明:该方法能够提高导航系统精度,使误差降低到惯导系统的8%左右。     

沉浸式虚拟训练中的虚拟手操控技术研究

将人手的自然动作加入沉浸式虚拟现实系统中,可有效提高系统的沉浸感和交互性,简便快捷、成本低廉的虚拟手建模及操控技术,有助于此项内容的推广应用;在对虚拟手骨骼结构模型分析的基础上,提出了基于Leap Motion的虚拟手控制方法和流程;以维修训练为任务背景,设计了四种用于虚拟手操作的手势规则,并对实现流程和关键技术进行了分析;构建基于Unity3D引擎的原型系统对虚拟手进行仿真验证,结果表明,基于Leap Motion的虚拟手操控简单、交互自然,可以满足虚拟现实系统交互需要。     

零视距地物长波红外特征场景仿真研究

为仿真地物长波红外场景图像,根据地表温度随时间变化的规律,并结合气象状况、背景材质、热特性参量、热状态等参数,在对太阳辐射、大气长波辐射、大气温度和地表热传导等影响地表温度变化的因素进行分析的基础上,建立了基于热平衡理论和热传导过程的方程。解算出多种常见地表一日之中的温度变化情况,并将其应用于由相同景物可见光纹理图像反演出的相应红外纹理图像中。在考虑景物表面自身发射、反射的辐射计算模型的前提下,生成了具有相似红外纹理细节的地表红外场景。结果表明,该方法可生成接近真实感的红外场景,有效地模拟仿真地物的长波红外特征。     

针对虚拟现实跟踪技术的快速滤波算法

提出一种用于实现手位手形跟踪的颜色滤波方法的计算模型。它可以取代虚拟现实中的数据手套,成为人机交互的工具。本方法是基于计算机视觉技术,模仿人眼的颜色视觉的特性,通过提取手的像素的特征颜色作为参考,对整个图像进行颜色滤波处理,从而将背景和干扰物体的像滤除,只保留手的像。然后经过平滑和轮廓提取处理,得到只含有手的轮廓的图像。颜色滤波算法能大大提高跟踪和识别的可靠性和实时性。     

A model for simulation of crowd behaviour in the evacuation from a smoke-filled compartment

The modelling of crowd evacuation from a building has been studied over the past decades. In this study, a numerical model based on cellular automaton is proposed to simulate the human behaviour termed “flow with the stream” in emergency evacuation from a large smoke-filled compartment. In the model, the smoke effect in the context of visibility is considered since visibility range can affect the human behaviour significantly. To simulate the reality that the smoke concentration in a fire compartment is not constant, the proposed model is developed to deal with the scenario in which the visibility range varies in the course of time. An empirical formula is incorporated into the proposed model to estimate the visibility range. The results of numerical tests show that the proposed model can also be used to investigate the effect of the number of guiders through case study.     

Simulation of pedestrian flow based on cellular automata: A case of pedestrian crossing street at section in China

One of the purposes of pedestrian studies is to evaluate the effects of a proposed program on the pedestrian facilities before its implementation. In order to evaluate the level of service (LOS) of a pedestrian facility, a microscopic model is built to simulate the process of pedestrian crossing street. Most of the existing models focus on the occupant evacuation flow in buildings; however, they are not appropriate for pedestrians in the traffic. According to the characteristics of pedestrian crossing street at signalized crosswalks, we build a model based on cellular automata. Both of the system size and cell size are coordinate with the reality. Depending on the contrast of three parameters of pedestrian flow between simulation data and the reality data, we found that this model is analogous to the real process of pedestrian crossing street at signalized sections. Finally, simulation and its results can provide guidance for evaluating the effects of pedestrian facilities before their implementation.     

A mass-conservative switching algorithm for modeling fluid flow in variably saturated porous media

Numerical solutions of flow equation in fluid content-based form or in fluid pressure head-based form are often tradeoffs between speed, accuracy, and convenience. The fluid-content based form can be solved quite rapidly with low CPU time and perfect mass balance. However, it cannot be used in saturated regions (as diffusivity function becomes infinite) and strictly becomes invalid in composite, layered, and real heterogeneous porous materials, due to singularity and discontinuity in fluid content profile. This formulation also gives misleading impression that gradient in fluid content causes the flow of fluid in porous materials, where in reality gravity and fluid pressure potential gradient produce the motion. The pressure head-based form, on the other hand, is more flexible but due to its highly nonlinear nature is much more time-consuming and produces poor global mass balance for dry initial conditions. Very fine spatial and temporal discretizations are needed to maintain mass balance property for these scenarios. The mixed form of the flow equation partially solves these issues as it maintains acceptable mass balance and is applicable to layered, heterogeneous, and composite fractured foundations. However, it is only applicable in unsaturated zones. In this study, a switching algorithm was proposed and implemented in which the mass conservative mixed form and the pressure head-based form were, respectively, used in the unsaturated and saturated zones of an initial-boundary value flow problem involving a variably saturated porous medium. The algorithm showed excellent agreement with a reference solution, obtained on a very fine spatiotemporal mesh. The simulator was then calibrated with several real-world large-scale experimental datasets. In all cases, the proposed algorithm exhibited close agreements with the experimental time–space series. The algorithm poses excellent mass balance property and can easily be used in both saturated and unsaturated regions without special treatment of fluid content discontinuities in heterogeneous and layered porous media. The proposed algorithm can also be extended to simulate multiphase and multidimensional flow problems.     

Dynamic simulation of locally inextensible vesicles suspended in an arbitrary two-dimensional domain,a boundary integral method

We consider numerical algorithms for the simulation of hydrodynamics of two-dimensional vesicles suspended in a viscous Stokesian fluid. The motion of vesicles is governed by the interplay between hydrodynamic and elastic forces. Continuum models of vesicles use a two-phase fluid system with interfacial forces that include tension (to maintain local “surface” inextensibility) and bending. Vesicle flows are challenging to simulate. On the one hand, explicit time-stepping schemes suffer from a severe stability constraint due to the stiffness related to high-order spatial derivatives in the bending term. On the other hand, implicit time-stepping schemes can be expensive because they require the solution of a set of nonlinear equations at each time step.     

Is a system's wave function in one-to-one correspondence with its elements of reality?

Although quantum mechanics is one of our most successful physical theories, there has been a long-standing debate about the interpretation of the wave function--the central object of the theory. Two prominent views are that (i) it corresponds to an element of reality, i.e., an objective attribute that exists before measurement, and (ii) it is a subjective state of knowledge about some underlying reality. A recent result [M.?F. Pusey, J. Barrett, and T. Rudolph, arXiv:1111.3328] has placed the subjective interpretation into doubt, showing that it would contradict certain physically plausible assumptions, in particular, that multiple systems can be prepared such that their elements of reality are uncorrelated. Here we show, based only on the assumption that measurement settings can be chosen freely, that a system's wave function is in one-to-one correspondence with its elements of reality. This also eliminates the possibility that it can be interpreted subjectively.     

用光晶格模拟狄拉克、外尔和麦克斯韦方程

相对论性量子力学波动方程,如狄拉克、外尔和麦克斯韦方程,是描述微观粒子运动的基石.最近的实验和理论研究表明,冷原子系统中几乎所有参数都可精确调控,因此冷原子系统被认为是实现量子模拟的理想平台,可以用来研究高能和凝聚态物理中的一些基本问题.本文介绍了设计原子光晶格哈密顿量的思路和方法,主要涉及激光辅助跳跃的理论.基于这些方法,物理学界提出了利用光晶格体系模拟相对论性量子力学波动方程,包括狄拉克、外尔和麦克斯韦方程等,并且预言了一些在基本粒子物理中很难观察到,但在冷原子体系可能观察到的物理现象.本文综述了国际上此领域的研究进展.     

The Evidence For and Against Various Theories of Light

The ether-wave theory of light, suitably modified, is fully supported by all known evidence. Further observation and analysis will be required to determine which of its several forms accurately represents reality. On the other hand, the tactile theory, the Newtonian corpuscular theory, the Ritz extinction theory, and the Special Theory of Relativity are not supported by the evidence.