利用深度图优化的自然光束体构建与绘制

为了模拟由散射形成的自然光束,提出基于深度图的光束体优化构建和实时绘制算法.首先把阴影体视作特殊的光束体参与计算,实现光束被物体遮挡的光照与阴影交错的效果;其次利用深度图在GPU中消除重叠的阴影体,降低填充率,优化复杂场景中多个阴影体与光束体相交情况的绘制效果.该算法将自然光束体构建中的一部分提前到预处理中完成,另一部分在GPU内核实现,提高了效率,再结合GPU上高效的散射计算绘制出动态光源产生的自然光束效果;此外,添加了大气中因散射可见的尘埃粒子的模拟,进一步增强了真实感.对动态光源下多个场景的实验结果表明,文中算法解决了散射光束中的遮挡与阴影问题,有效地模拟了光影交错的综合效果,并具有实时性.     

真实感海底场景的实时绘制

海底场景的仿真对于动画游戏、海洋勘测、航海驾驶、灾害救援等有着重要的应用意义。由于涉及更加复杂的海水与光线、地形等的交互,海底场景中的复杂刻蚀、光束、散射等效果是实时绘制海底场景的难点。基于海底环境的物理机理,首先提出基于线框绘制模式结合高斯滤波的方法来模拟精细的刻蚀效果,采用贴近真实的散射相位函数计算海底散射来模拟水下海水颜色,采用柱面组织的光束算法来模拟海底光线效果,并进一步给出一种基于空间划分的海底场景绘制优化方案,最后基于GPU加速技术实现了不同情况下的真实感海底场景的实时绘制。     

多视角实时大气渲染算法研究

提出了一种在把瑞利散射和米氏散射列入考虑的情况下,从地平面到外太空的任何视角实时渲染大气的算法.该算法能够呈现出许多光的散射现象,例如目光和曙暮光天空的颜色和航空视角等所有视角下的光线方向,或者地面和山脉的阴影在大气里形成的光束,并展示了运用一个适应GPU的算法在几秒钟之内预处理完成这些数据.这个预处理的数据使得我们能够在考虑到地面上阴影和自然光束的情况下,评估光线传输方程在固定时间内的运行时间,从而克服了常规方法在实时性方面的瓶颈,并且有良好的显示效果,在航空仿真等领域可以得到广泛应用.     

基于流场的流动水体仿真

为了逼真地模拟自然河道中的流动水体,提出了一种基于流场的流动水体仿真方法.应用流体力学原理,实时计算稳定水流的速度,构建河流的流场,然后运用流场驱动并约束满足泊松碟分布的块状纹理在河道内移动,通过对块状纹理进行混合与渲染,构成了一种自然水体的动态流动效果.水面渲染采用GLSL (opengl shading language)着色器进行渲染,实际运用GPU可编程渲染管线进行图形计算,减少了CPU的实时运算量,提高程序的整体效率,实践表明,应用以上方法可以有效地模拟流域内的动态流动水体.     

GPU在实时阴影绘制中的应用

实时阴影在增强三维场景真实感方面起着非常重要的作用。阴影体算法是实时阴影绘制中效果非常理想的一种方法。但是随着场景复杂度的增加,该算法计算量比较大,将导致绘制效率的降低。另一方面,随着可编程GPU技术的发展,GPU的渲染速度远远大于CPU,为提高三维场景的渲染效率提供了更大的空间。在此基础上,介绍了一种在GPU上生成阴影体的方法,加速实时阴影绘制。利用图形硬件的图形处理单元（GPU）的运算能力和可编程性,将生成阴影体的大量计算从CPU转移到GPU,从而有效地提高实时阴影的绘制效率。     

基于GPU的网格模型平滑阴影的实时绘制

为了绘制边界平滑的阴影效果,增强虚拟场景的真实感,提出一种基于GPU的平滑阴影实时绘制的算法.该算法通过重新定义阴影轮廓线得到网格模型上的精确阴影轮廓边界,利用阴影体算法获得边界平滑的阴影区域;阴影轮廓线的计算生成及阴影Ⅸ域的绘制等过程利用GPU进行加速,可以实时绘制物体的阴影区域.实验结果表明,文中算法生成的阴影边界比已有算法更加平滑,并且在效率方面强于其他算法.     

利用反向投影实现的实时软阴影映射算法

基于阴影映射算法,提出一种利用反向投影实现的实时软阴影的新算法。算法对每个光源都产生对应的阴影图,使用阴影图作为对场景的离散化表示,引入可见因子来计算场景点的亮度信息,并采用GPU片元着色、层次阴影图、自适应精度等方法加速阴影渲染。实验表明,算法非常适合于实时渲染复杂、动态的场景,可以很好地处理遮挡物的融合,并且很容易在可编程图形硬件上实现。     

基于GPU的大规模波动草叶实时渲染技术

本文通过有效利用图形硬件图形处理单元(GPU)的运算能力和编程性,将大量计算从CPU中分离出来,实现了自然逼真而且高效的大规模波动草叶的实时渲染.利用GPU的顶点程序进行草叶的运动计算,利用GPU的片元程序进行静态阴影的计算.本文技术由OpenGL结合Cg编程实现,达到了自然逼真的渲染效果和较高的渲染效率.     

基于高度梯度图的微结构表面阴影绘制方法

针对带有微结构表面的几何模型建模困难、模型复杂度高,导致现有阴影绘制算法效率差的问题,提出一种基于高度梯度图的阴影实时绘制方法.为实现实时渲染,首先给出一种微结构描述模型,通过将微结构高度场映射到低精度模型表面来描述复杂微结构模型;然后定义微结构高度梯度图,并据此构建可见点的局部最高点集合.在阴影绘制阶段,通过实时计算梯度空间局部最高点集合遮挡来生成细节阴影.整个阴影绘制算法在图像空间完成,较好地利用了延迟着色的思想和GPU并行计算的特点,可呈现可变形物体在动态场景中的自阴影效果.     

可扩展的实时自然景物模拟算法

针对传统的粒子系统实时仿真存在只能针对单一自然景物模拟、计算耗时、图像不真实、算法复杂等问题,提出了一种基于粒子系统和图形处理器（GPU）加速通用可扩展的自然景物模拟算法。在该算法中,粒子的物理运动计算过程和渲染阶段完全由CPU转移至GPU,可以增加粒子数量和提高渲染速度;同时,在渲染过程中,可以较好地利用硬件支持的粒子图技术来改善渲染中粒子的外表,选择不同纹理,从而能够较方便地模拟不同的自然景物。最后,在GPU上实现了雪花、喷泉、烟花、瀑布等模拟,算法充分利用了GPU的多通道并行处理性和可编程性,提高了自然景物模拟的实时性,可运用于虚拟现实系统。     

Realistic real-time rendering of light shafts using blur filter: considering the effect of shadow maps

The ray marching method has become the most attractive method to provide realism in rendering the effects of light scattering in the participating media of numerous applications. This has attracted significant attention from scientific community. Up-sampling of ray marching method is suitable for rendering light shafts of realistic scenes, but suffers of consume a lot of time for rendering. Therefore, some encouraging outcomes have been achieved by using down-sampling of ray marching approach to accelerate rendered scenes. However, these methods are inherently prone to artifacts, aliasing and incorrect boundaries due to the reduced number of sample points along view rays. This research proposes a realistic real-time technique to generate soft light shafts by making use downsampling of ray marching in generating light shafts. The bilateral filtering is then applied to overcome all defects that caused by downsampling process to make a scene with smoothing transition while preserving on the edges. The contribution of this technique is to improve the boundaries of light shafts taking into account the effect of shadows. This technique allows obtaining soft marvelous light shafts, having a good performance and high quality. Thus, it is suitable for interactive applications.

     

Precomputed Atmospheric Scattering

We present a new and accurate method to render the atmosphere in real time from any viewpoint from ground level to outer space, while taking Rayleigh and Mie multiple scattering into account. Our method reproduces many effects of the scattering of light, such as the daylight and twilight sky color and aerial perspective for all view and light directions, or the Earth and mountain shadows (light shafts) inside the atmosphere. Our method is based on a formulation of the light transport equation that is precomputable for all view points, view directions and sun directions. We show how to store this data compactly and propose a GPU compliant algorithm to precompute it in a few seconds. This precomputed data allows us to evaluate at runtime the light transport equation in constant time, without any sampling, while taking into account the ground for shadows and light shafts.     

Neural net based torque sensor using birefringent materials

Photoelasticity can be used to accurately measure surface strains or stresses in a part or structure. In this paper we describe the use of a photoelastic transducer and neural net image processing to estimate the torque of stationary and rotating shafts. A strain sensitive (photoelastic) plastic cylinder is attached to the shaft and illuminated by polarized light. As the shaft torque varies the photoelastic plastic displays the corresponding shaft strain as a two-dimensional fringe pattern when viewed through an optical polarizer. The strain that causes this observed optical pattern is a complex function of the torque applied to the shaft. In this paper, we describe the use of neural net image processing to estimate this function to realize an optical torque sensor. A CCD camera/image processing system was used to acquire and process the optical patterns. A neural net torque estimator was trained with these fringe patterns and tested against a laboratory strain gauge torque sensor. Our experiments show that the neural net torque estimator can accurately estimate (to within a few percent) the applied torque for both static and slowly rotating (<20 rpm) shafts.     

LD􀀁PSD􀀁直角棱镜同轴度测量技术研究

半导体激光准直光源(LD)和光斑位置传感器(PSD)固定在从动轴上,直角棱镜固定在主动轴上.并根据该结构建立了测量过程的数学模型,讨论了所存在的主要系统误差以及为消除系统误差的影响所应采取的措施.结论:若激光器不存在安装倾斜误差,PSD传感器也不存在安装平移误差,则在任意两个转角位置进行测量均可得到4个同轴度偏差参数,否则要在任意四个转角位置进行测量方能消除激光器和PSD传感器的安装误差对测量结果的影响.     

双LD-双PSD同轴度测量技术研究

介绍了双LD-双PSD同轴度测量仪的结构：在主、从动轴上相向各固定一支半导体激光准直光源（LD）和光斑位置传感器（PSD）,激光束在对方的PSD光敏面上形成测量光斑;据该结构建立了测量过程的数学模型,讨论了所存在的主要系统误差以及为消除系统误差的影响所应采取的措施;结论：若激光器不存在安装倾斜误差,PSD传感器也不存在安装平移误差,则在任意一个转角位置进行测量可得到4个同轴度偏差参数,否则要在任意两个转角位置进行测量方能消除激光器和PSD传感器的安装误差对测量结果的影响。     

A pocket calculator program for some simple vibration problems

An earlier buckling program [1] has been adapted to give the natural frequencies and mode shapes for free axial vibrations of bars, transverse vibrations of taut wires and torsional vibrations of shafts. The cross section of the bar, wire or shaft can vary at intervals along its length, and so can its distributed mass (or inertia for the shaft), which is allowed for in a classically exact way. Lumped masses (or rotors for shafts) and spring supports can be attached anywhere along the length, and the ends can be free or clamped. The program just fits on a Hewlett-Packard HP-25 pocket calculator, which can store only 49 program steps and 13 numbers.     

煤矿新建井筒安全监测系统的设计与应用

针对煤矿井筒建设与生产过程的特点,提出了一种煤矿新建井筒安全监测系统的设计方案,详细介绍了系统结构组成、基本功能设计及其在某煤矿井筒中的应用情况。该系统基于煤矿局域网,集自动化、信息、计算机、网络、通信技术于一体,可监视各井筒的安全状态,收集有关的安全参数,保证了各井筒的安全,实现了煤矿各部门之间的信息交换。     

A Rendering Algorithm for Discrete Volume Density Objects

We present a new algorithm for simulating the effect of light travelling through volume objects. Such objects (haze, fog, clouds…) are usually modelized by voxel grids which define their density distribution in a discrete tridimensional space. The method we propose is a two-pass Monte-Carlo ray-tracing algorithm that does not make any restrictive assumptions neither about the characteristics of the objects (both arbitrary density distributions and phase functions are allowed) nor about the physical phenomena included in the rendering process (multiple scattering is accounted for). The driving idea of the algorithm is to use the phase function for Monte-Carlo sampling, in order to modify the direction of the ray during scattering.     

Single and multi-purpose optimization of vibrating Timoshenko shafts

Optimization of transversely vibrating shafts with respect to eigenfrequencies, with constraints on the design variables, has been almost fully investigated in recent years - as far as the Bernoulli-Euler equation of motion is concerned. In the present paper, Timoshenko's equations are applied, i.e. the effects of shear and rotational energy are taken into account in the calculation of the transverse vibration modes and frequencies of the shaft. This is done via a finite element formulation. The sensitivity analysis of the transverse vibration frequencies is based on analytical differentation of the stiffness- and mass matrices. The computer program developed provides the user with an option to suppress the Timoshenko effects, such that the analysis can be carried out within the Bernoulli-Euler theory as well and then comparisons can be made. Torsional vibrations of shafts are also considered, computing the vibration modes and natural frequencies by means of a finite difference approach. Again, the sensitivity analysis is carried out analytically. Objective functions and behavioral constraints are selected from fundamental frequencies, higher order frequencies and differences between adjacent frequencies of the vibrational types considered. The cross-sectional area function of the shaft is used as the design variable and the total volume of structural material is assumed to be given, along with some sizing constraints. The shafts considered may be equipped with non-structural disks and flexible supports. Several examples are presented and some notable differences between optimized Timoshenko and Bernoulli-Euler shafts derived.