Experimental Study of the Earthquake Recurrence Period and the Trend of Post-Seismic Development

In order to study the earthquake recurrence and the characteristics of earthquake series, rupture tests of rock samples and plexiglass samples were made. On rock samples, a number of acoustic emission (AE) and strain measuring points were deployed; the load was one side direct shear. The variation characteristics of AE and strain at different detecting points around the extra large fracture were observed and studied. On plexiglass samples, a series of inclined cracks were prefabricated by a small-scale compressive testing machine. The samples were then loaded on a shockproof platen, when the samples were loaded, the stress intensity factor (SIF) was determined by the laser interferometric technique and shadow optical method of caustics. The fracture conditions such as material toughness around the extra large fracture were also studied. From those experimental results and the theory of fracture mechanics, the earthquake recurrence period and the trend of post-seismic development were studied.

An Experimental Study on Flame Propagation in Cornstarch Dust Clouds

Following the quantitative determination of dust cloud parameters, this study investigates the flame propagation through cornstarch dust clouds in a vertical duct of 780 mm height and 160 x 160 mm square cross section, and gives particular attention to the effect of small scale turbulence and small turbulence intensity on flame characteristics. Dust suspensions in air were produced using an improved apparatus ensuring more uniform distribution and repeatable dust concentrations in the testing duct. The dispersion-induced turbulence was measured by means of a particle image velocimetry (PIV) system, and dust concentrations were estimated by direct weighing method. This quantitative assessment made it possible to correlate observed flame behaviors with the parameters of the dust cloud. Upward propagating dust flames, from both closed/open bottom end to open/closed top end of the duct, were visualized by direct light and shadow photography. From the observation of propagation regimes and the measurements of flame velocity, a critical value of the turbulence intensity can be specified below which laminar flame propagation would be established. This transition condition was determined to be 10 cm/s. Laminar flames propagated with oscillations from the closed bottom end to the open top end of the testing duct, while the turbulent flames accelerated continuously. Both laminar and turbulent flames propagated with steady velocity from the open bottom end to the closed top end of the duct. The measured propagation velocity of laminar flames appeared to be in the range of 0.45-0.56 m/s, and it was consistent with the measurements reported in the literature. In the present experimental study, the influence of dust concentration on flame propagation was also examined, and the flame propagation velocity was found weakly sensitive to the variations in dust concentration. Some information on the flame structure was revealed from the shadow records, showing the typical heterogeneous feature of the dust combustion process.

Analytic expression of the diffraction of a circular aperture

Recurring to the characteristic of Bessel function, we give the analytic expression or the Fresnel diffraction by a circular aperture, thus the diffractions on the propagation axis and along the boundary of the geometrical shadow are discussed conveniently. Since it is difficult to embody intuitively the physical meaning from this series expression of the Fresnel diffraction, after weighing the diffractions on the axis and along the boundary of the geometrical shadow, we propose a simple approximate expression of the circular diffraction, which is equivalent to the rigorous solution in the further propagation distance. It is important for the measurement of the parameter or the beam, such as the quantitative analysis of the relationship of the wave error and the divergence of the beam, In this paper, the relationship of the fluctuation of the transverse diffraction profile and the position of the axial point is discussed too. (c) 2005 Elsevier GrnbH. All rights reserved.

Sleeping sites of Rhinopithecus bieti at Mt. Fuhe, Yunnan

Data on sleeping site selection were collected for a group of black-and-white snub-nosed monkeys (Rhinopithecus bieti; around 80) at Mt. Fuhe, Yunnan, China (99degrees20'E, 26degrees25'N, about 3,000 m asl) from November 2000 to January 2002. At the site mainly three vegetation types were present in an elevation-ascending order: deciduous broad leaf forest, mixed coniferous and broad leaf forest, and dark coniferous forest. In addition, bamboo forest presented in areas burned in 1958. Sleeping sites (n = 10) were located in the coniferous forest, where trees were the tallest, bottommost branches were the highest, the diameter of crowns was the second largest, and the gradient of the ground was the steepest. Monkeys usually kept quiet during entering and staying at a sleeping site. The site choice and the quietness may be tactics to avoid potential predators. In the coniferous forest, however, monkeys did not sleep in the valley bottom where trees were the largest, but frequently slept in the middle of the slope towards the east/southeast, in the shadow of ridges in three other directions, to avoid strong wind and to access sunshine; in winter-spring, they ranged in a more southern and lower area than in summer-autumn. These may be behavioral strategies to minimize energy stress in the cold habitat. Monkeys often slept in the same sleeping site on consecutive nights, which reflected a reduced pressure of predation probably due to either the effectiveness of anti-predation through sleeping site selection, or the population decline of predators with increasing human activities in the habitat. The group's behavioral responses to interactive and sometimes conflicting traits of the habitat are site-specific and conform to expectations for a temperate zone primate.

伪柔和阴影算法的研究

在计算机图形学领域，阴影的绘制一直是一个热点的研究内容，它对增强场景的真实感有着非常重要的意义。完全物理正确的柔和阴影绘制通常需要耗费大量的时间，于是研究者们提出了各种阴影的近似方法。 本文首先对阴影绘制算法的研究发展历程作了概要性的介绍，特别针对近几年国内外提出的伪柔和阴影算法做了详细的介绍和分析。这些算法主要是根据点光源得到的阴影图[1]和轮廓边进行近似软影绘制的算法，不仅绘制效率高，而且画面真实感比较强。 随着GPU的不断发展，其强大的并行计算能力以及新的可编程着色器，特别是几何着色器[2]的出现，使得基于阴影图的阴影绘制算法中可以加入和场景几何相关的一些计算。本文提出了一个深度剥离[3]和GPU相结合的实时近似软影绘制算法，该算法利用了GPU强大的可编程能力，利用场景物体轮廓信息生成半影图，以此得到伪柔和阴影。实验结果表明本算法是可行的。

阴影实时绘制技术研究

阴影是自然界中一种普遍存在的自然现象，它能够提供场景数据之外的感知信息，因此阴影的实时绘制是真实感技术研究的主要内容之一。本文围绕阴影的实时绘制技术这一主题展开，主要研究内容包括以下几个方面： （1）预滤波阴影贴图反走样技术研究 方差阴影贴图算法是一种基于概率方法的预滤波算法，它本身存在着比较严重的光渗现象。我们采用高次幂函数对深度进行非线性映射，按照这一深度根据方差阴影贴图算法求得一个可见性数值，并与原始的方法求得的可见性数值求取一个最小值，作为最终的可见性，从而能够很好地消除光渗现象。同时，我们针对指数阴影贴图算法的不稳定性，将其与方差阴影贴图算法相结合，采用混合的方法消除二者存在的缺陷。 （2）基于GPU的近似软影绘制 通过对阴影图算法进行扩展，提出了一种完全基于GPU的近似软影实时绘制算法。第一遍与传统的阴影图算法类似，从光源中心计算场景的深度图。第二遍采用几何着色器提取物体的轮廓边，同时在轮廓边上生成新的几何图元，利用硬件自动插值功能，向外绘制线性近似半影图，并根据第一遍得到的深度图在像素着色器中对背面轮廓形成的半影区进行剔除；对于重叠的半影区，通过设定片元的伪深度值利用硬件进行自动融合；这一遍是完全在GPU中运行的，与其它方法不同的是我们只需要一遍绘制即可得到近似半影区纹理。第三遍分别查询深度图和半影图，以确定场景的本影区以及半影区中像素的亮度，从而得到面光源照射下场景的近似软影效果。 （3）基于预计算辐射率传输的近似软影绘制 基于预计算辐射率传输的方法，提出了一种基于圆盘近似的全动态场景柔和阴影实时绘制算法。该方法首先采用圆盘对物体表面进行近似，对每个圆盘形成的遮挡使用椭圆遮挡域表示，预计算过程中根据每个椭圆遮挡域参数进行采样并用球面调和系数表示，同时将其转换为对数，对采样数据采用主元分析方法进行压缩；在绘制的过程中，根据圆盘的法线、半径以及和阴影接收点之间的相对位置确定其椭圆遮挡域的值，然后采用球面调和指数算法进行累积。该方法将圆盘近似算法和球面调和指数算法相结合，能够灵活地描述柔性物体以及较薄的物体形成的柔和阴影。 （4）近似全局光照算法研究 提出了一种混合的近似全局光照算法，对原始模型采用卷积软影图算法计算直接光照，而间接光照采用简化模型计算。在直接光照计算过程中同时生成光源视点的场景的颜色、法线以及世界坐标位置纹理，利用准随机采样的方法从这三幅纹理中获得间接光照采样点，称之为像素光源，然后利用抛物面阴影贴图算法对简化模型计算间接光照；最后将直接光照和间接光照融合获得最终结果。为了进一步提高精确性，仅用简化模型生成阴影贴图，间接光照绘制的过程中采用Deferred Shading的方式进行计算，将视域中可见像素的位置和法线保存在纹理中，间接光照计算直接对纹理中的像素进行。该方法能够较好地处理动态场景的近似全局光照效果。

方差阴影图中的光渗现象消除算法

方差阴影图算法使用概率的方法计算像素被遮挡的上限概率,通过对深度图滤波的方法来有效地减少阴影图算法中的走样问题,但在深度比较复杂的场景中方差阴影图算法会出现光渗现象,即在应该是阴影的区域却有了亮度.文中使用最小-最大阴影图来辅助消除方差阴影图中的光渗现象,在对深度纹理进行滤波的同时生成一个最小-最大阴影图;在实时绘制场景时,利用最小-最大阴影图来辅助判断当前片元是否完全处在阴影区域内部,由此生成更真实、更准确的阴影.该算法可以很容易地添加到已有的方差阴影图算法的片元处理程序中,并且不会对原有阴影的柔和边界以及绘制的帧率产生影响.

基于GPU的近似软影实时绘制

通过对阴影图算法进行扩展,提出一种完全基于GPU的近似软影实时绘制算法,它是一种3遍算法:第一遍从光源中心计算场景的深度图;第二遍采用几何着色器提取物体的轮廓边,同时在轮廓边上生成新的几何图元,利用硬件自动插值功能向外绘制线性近似半影图,并根据第一遍得到的深度图在像素着色器中对背面轮廓形成的半影区进行剔除;对于重叠的半影区设定片元的伪深度值,利用硬件进行自动融合.第三遍分别查询深度图和半影图,确定场景的本影区以及半影区中像素的亮度,从而得到面光源照射下场景的近似软影效果.

Total-internal-reflection fluorescence microscopy with W-shaped axicon mirrors

A scheme based on a W-shaped axicon mirror device for total-internal-reflection fluorescence microscopy (TIRFM) is presented. This approach combines the advantages of higher efficiency compared with traditional TIRFM, adjustable illumination area, and simple switching between wide-field and TIRF imaging modes. TIRF images obtained with this approach are free of shadow artifacts and of interference fringes. Example micrographs of fluorescently labeled polystyrene beads, of Convallaria majalis tissue, and of Propidium-iodide-labeled Chinese hamster ovary cells are shown, and the capabilities of the scheme are discussed. (C) 2010 Optical Society of America

基于阴影图的伪柔和阴影绘制算法

本文提出一种基于阴影图预滤波的伪柔和阴影实时绘制算法。 阴影是真实世界中的一种十分普遍的物理现象，它不但可以反映物体自身的形状，而且可以反映空间中物体与物体之间的相对位置关系，对于增加场景的真实感具有非常重要的意义。因而阴影的实时绘制一直都是计算机图形学中一个重要的研究方向。 方差阴影图算法是对传统阴影图算法的一个重要改进。传统的阴影图算法虽然有着易于实现、硬件支持好、与场景复杂度无关等优点，但是由于它是对场景的离散表示，所以会出现各种各样的走样问题。方差阴影图算法是使用概率的方法计算片元被遮挡的上限概率，它可以通过对深度图滤波的方法来有效地减少阴影图算法中的走样问题。方差阴影图算法克服了传统阴影图算法中边界走样问题，可以生成比较柔和的边界。但是，这种概率方法在深度比较复杂的场景中会出现光渗现象，即在原本应该是阴影的区域却有了亮度。本文使用最小-最大阴影图来辅助消除方差阴影图中的光渗现象。算法在对深度纹理进行滤波的同时生成一个最小-最大阴影图；在从视点实时绘制场景时，利用最小-最大阴影图来辅助判断当前片元是否完全处在阴影区域内部。通过将最小-最大阴影图和方差阴影图相结合，本文的算法可以快速生成伪柔和阴影。

Undercut structures fabricated by microtransfer printing combined with UV exposure and their applications

In this paper. the undercut structures were fabricated by microtransfer printing of metal films on the surface of photoresist combined with UV exposure and photoresist film developing. The patterned metal films were used as mask to realize the selective UV exposure of photoresist Firstly. The undercut structures. which consist of the top metal films and the patterned bottom photoresist, formed in the subsequent developing process because of the lateral dissolving of photoresist at the edge of the unexposed regions. The method proposed in this paper has wider tolerance to the changing of the patterning parameters. but, without effect on patterning resolution since the metal film was used as the top layer. The undercut Structures were Used as separators to pattern passive-matrix display of organic light-emitting diodes (OLEDs). No visible difference of the device performance was observed Compared with the OLEDs patterned by the shadow mask.

用光学阵列传感器的机器人物体分类系统

本文介绍用光学阵列传感器的机器人物体分类系统。传感器直接安装在机器人的两个手指上。被抓物体的阴影通过光导纤维传到安放在“安全区”的光敏元件上。计算机识别物体的轮廓后命令机器人抓握物体,并把它运送到指定的地点从而达到物体分类的目的。