Protein engineering of alcohol dehydrogenases; effects of amino acid changes at positions 93 and 48 of yeast ADH1

By protein engineering we have investigated changes to two aminoacid residues (Trp93 and Ser48) in the substrate pocket of yeastalcohol dehydrogenase 1. Upon changing Thr48 to serine we producedan enzyme which has markedly greater activity towards aliphaticalcohols with chain length up to 8, together with a generalincrease in catalytic activity (V/K). Changes at position 93were less pronounced, with the Phe enzyme being more activethan the parent towards the range of alcohols but with the alanineenzyme showing very little difference from the wild-type. Enzymeswith the double changes at 48 and 93 showed increased activitytowards alcohols with 3–8 carbons but the increases werenot additive over the single changes. The enzymes with changesat the two positions would metabolize both stereoisomers of2-octanol whereas the parent ADH would attack only one of them.None of the engineered enzymes would attack cyclohexanol oraromatic alcohols. The results are in general agreement withthe prediction that reducing the size of amino acids in thesubstrate pocket would enhance the ability to oxidize alcoholslarger than ethanol.     

Numerical analysis of particle mixing in a rotating fluidized bed

This paper describes the numerical analysis of particle mixing in a rotating fluidized bed (RFB). A two-dimensional discrete element method (DEM) and computational fluid dynamics (CFD) coupling model were proposed to analyze the radial particle mixing in the RFB. Spherical polyethylene particles (Geldart group B particles) were used as model particles under the assumptions that they were cohesionless and mono-disperse with their diameter of 0.5 mm.The validity of the proposed model was confirmed by the comparison between the calculated degree of particle mixing and the experimental one, which was obtained by measuring the lightness of the recorded image taken by a high-speed video camera. Effects of the operating parameters (gas velocity, centrifugal acceleration, particle bed height, and vessel radius) on the radial particle mixing rate were numerically analyzed. The radial particle mixing rate was found to be strongly affected by the bubble characteristics, especially by the bubble size. The mathematical model for the rate coefficient of particle mixing as functions of operating parameters was empirically proposed. The radial particle mixing rate in a RFB could be well correlated by the three dimensionless numbers: dimensionless acceleration (Ac), bubble Froude number (Fr_b), and dimensionless radius on the surface of particle bed (β_s).     

军事训练能力要素研究

军事训练能力是作战能力生成发展的重要基础和可靠保证。系统地研究了军事训练能力的概念和构成要素,提出了训练本领集成和训练力量体现的军事训练能力新概念,分析了军事训练能力的3个要素和各要素之间的相互作用及运行过程。     

一种改进的文档图像倾斜检测算法

在分析现有倾斜检测算法基础上,提出一种改进的文档图像倾斜检测算法。该算法首先对文档图像进行图像增强、去噪和二值化处理,然后提取边界线、计算每条边界线的角度,最后计算加权平均得到文档图像的倾斜角度。通过在计算每条边界线的角度过程中引入LMS算法。改进的算法提高了原算法的精确度、稳定性和适应性。实验表明,改进的方法有效可行。     

基于图像反馈的标靶面与光轴垂直调节系统

介绍了摄像机标定的内部参数,阐述了标靶面与光轴垂直可以简化标定的过程.当标靶面与光轴垂直时,利用标靶面上特征圆在CCD成像面上成像也是圆的特点,提出了基于图像反馈的标靶面与光轴垂直的调节方法.一般情况下,标靶面与光轴不垂直,标靶面上特征圆在CCD成像面上成像为椭圆.通过图像处理和最小二乘椭圆拟合法计算出成像椭圆的长短轴及偏角,设计了调整控制策略.实验结果表明,该方法是有效可行的.     

图像重建的统计自适应子集算法

有序子集算法大大提高了图像重建迭代算法的收敛速度.增加子集的数目可以加速收敛,但图像质量会由于子集内缺少统计信息而下降.提出了一种基于假设检验的子集划分方法,建立了检验统计量,给出了算法的迭代公式.该算法可以根据用户定义的显著性水平,在每次迭代时自动调节子集的个数,生成含有相同统计信息量的子集.实验结果表明:该方法可以在少数次迭代后得到较高质量的重建图像.     

结合PCM和MIE的多模态图像配准方法

针多模态图像配准问题,提出了一种基于相位一致性模型（PCM）和互信息熵（MIE）的配准方法.通过相位一致性模型同时提取多模态图像的角点和边缘特征,在边缘图上取角点的邻域,依据邻域间互信息熵的最小值在估计区域搜索匹配特征,利用RANSAC算法去除错配,进而确定待配准图像间的变换参数.实验表明：该方法达到了像素级配准精度、求解稳定,对多模态引起的非线性灰度变化、光照变化、噪声等都具有较强的鲁棒性;计算精度较基于同类特征的配准方法高,角点、边缘等几何特征综合运用在多模态图像配准中效果良好.     

一种用于图像篡改定位的半脆弱数字水印算法

设计并实现了一种用于数字图像内容认证的半脆弱水印方案。算法中,图像块的水印选择Logistic映射作为混沌系统从该图像块本身产生一系列混沌序列,作用于水印的嵌入;每个图像块产生的水印按照Torus,自同构映射嵌入另一个映射块的LSB(Least Significant Bit),对应关系通过密钥来确定,这样建立起图像块之间的相关性;利用混沌对初值极端敏感性的特点,能够定位检测对加入水印图像的篡改,且水印提取不需要原始图像。实验结果表明,嵌入水印后的图像的视觉好;算法具有图像内容局部篡改检测的有效性、敏感性以及良好的篡改定位能力。     

模拟尾流气泡幕的灰度图像分析

为了对舰船尾流产生的大量气泡进行实时探测,对气泡的密度分布进行了分类判别,利用灰度图像直方图统计矩法,分析了随着压强增大,灰度直方图的峰值由较暗向较亮方向移动.应用图像处理技术对不同压强下模拟尾流气泡幕的统计特性参数(均值、三阶矩、标准偏差、归一化、一致性、熵)进行分析,研究结果表明:随着压强的增大气泡幕图像的均值递增说明平均灰度级在递增;三阶矩呈递减状态反映了图像灰度直方图的"歪斜程度";在其他4个参数的曲线中0.025 MPa所对应的值说明了所采集的气泡幕图像的纹理特征比0.02 MPa,0.03 MPa,0.035 MPa压强下的要粗糙;所提取的六个特征参数准确的表征了不同压强下气泡幕的特征.     

基于神经网络监督控制的拥塞控制算法研究

提出了一个基于神经网络控制的主动队列管理（AQM）算法;研究了TCP/AQM拥塞控制系统的可逆性,并利用一种神经网络监督控制结构进行了AQM算法的设计。算法由一个三层前馈结构的神经网络控制器（neural network controller,NNC）和一个反馈控制器（feedback controller,FC）组成。NNC作为一个前馈控制器,通过FC产生的教师信号进行学习,以建立被控对象的逆动力学模型。仿真结果表明,提出的算法与PI（proportional-integral）算法相比,无论在瞬态性能