基于主曲率和主方向的多尺度视网膜血管骨架提取方法

为了减少视网膜血管骨架提取过程中低对比度血管的漏检和误检数量,提出了一种基于主曲率和主方向的多尺度视网膜血管骨架提取方法。首先,分别提取视网膜图像中每个像素点在多尺度高斯滤波后的主曲率和主方向;其次,在每个尺度下分别提取最大主曲率方向上的局部最大值点,并通过曲率阈值筛选出高对比度血管中心像素点作为种子点;然后,利用最小主方向对低对比度血管进行骨架追踪和标记;最后,对多个尺度下提取的血管骨架进行融合。方法分别在DRIVE训练数据集、DRIVE测试数据集和STARE数据集上进行了测试,漏检数量分别为89、97、106,误检数量分别为99、101、122。实验结果表明,该方法能够提取出低对比度的细小血管骨架,但对于对比度在3个灰度级以内的细小血管存在少量漏检,对于与血管粘连的高对比度细条状纹理和病灶干扰存在少量误检。     

Dynamic characteristics of NC table with SVD

This paper employs the SVD (singular value decomposition) method to study dynamic characteristics of a numerical control (NC) table. Acceleration signals of the NC table at three directions are tested; the singular spectrum of the signals is acquired with SVD; principal components of the signals are found out; dynamic characteristics of the signals and their contributing factors are studied by extracting dynamic characteristics of principal components; and signals and principal components are quantitatively analyzed by calculating signal energy. Results indicate that signal characteristics of the previous two principal components are apparent, based on which dynamic characteristics of chaotic signal can be extracted. Signal at the perpendicular direction of the table is significantly correlated with that at the horizontal motion direction, which indicates that they are excited from the same vibration source. However, signals perpendicular to each other in terms of the motion direction at the horizontal level are rarely correlated; the total signal energy is maximum at the motion direction, minimum at the horizontal non-motion direction, and medium at the perpendicular non-motion direction. Bending vibration of the lead screw at the perpendicular direction is far more violent than that at the horizontal direction.     

3-UCR三自由度并联机构瞬态运动学研究

针对一种特殊3-UCR并联机构,基于反螺旋理论与虚拟机构法对其瞬态运动特性进行了分析,利用支链与动平台的一阶影响系数矩阵之间的关系,建立了此并联机构在螺旋坐标系下的瞬态运动数学模型,得出了一阶影响系数矩阵只与机构输入参数、机构构型有关的结论,同样此建模方法适用于其它欠自由度并联机构的瞬态运动特性分析。通过此数学模型,研究得到任意瞬态的螺旋系节距,表明该机构具有两转动一平动的瞬态运动特性,并利用动静平台平行位姿时的具体数值实例,验证了此瞬态运动特性符合前述分析结果。     

基于PCA与Wilks准则的电子鼻酒类鉴别方法研究

在电子鼻的模式识别方法中,主成分分析（PCA）是常用的方法之一。然而,主成分分析在计算过程中消除了各变量（对应于电子鼻的各测量传感器）问的相关性,这与传感器阵列的交叠感应特性相悖,致使时常无法正确鉴别多组分物品的类别。本文将Eilks准则引入主成分分析中,解决了酒类鉴别中主成分主轴向量的选择问题,实现了3种不同种类酒的正确鉴别,突破了纯粹的主成分分析模式。同时,指出了在用电子鼻鉴别多组分物品的种类时,主成分主轴的选择并非完全按照主成分贡献率的大小来确定的,这为今后的相关研究提供了一些有益的参考。     

稀薄效应对动压气体轴承静动特性的影响

为研究稀薄效应对微小间隙下动压气体轴承静动态特性的影响,分别采用连续模型、一阶滑移模型以及在任意克努森数下都与线性玻尔兹曼方程解有较高吻合的WU新滑移模型,基于有限体积法建立考虑稀薄效应的静动态Reynolds方程,给出3种模型下轴承静态承载力与偏心率、轴颈扰动频率与轴承动特性系数的变化关系。数值分析结果表明:随偏心率的变化,连续模型预测的静态承载力最大,一阶滑移次之,WU新滑移模型预测的值最小;随着扰动频率的增加,考虑滑移模型计算的主刚度系数和主阻尼系数均有减小的趋势,且WU新滑移模型计算的主刚度系数和主阻尼系数明显低于连续模型和传统的一阶滑移模型。传统连续模型和一阶滑移模型过高地估计了轴承的静动特性系数,WU新滑移模型计算的结果更为准确。     

VOLUME DEFECT FATIGUE FAILURE OF CERAMIC BALLS UNDER ROLLING CONDITION

A newly developed pure rolling fatigue test rig with three contact points is used to test the rolling contact fatigue properties of silicon nitride ceramic balls. Ball surfaces are examined after failure with optical microscopy and scanning electron microscopy. The failure cause,fatigue phenomenon and mechanics are analyzed. The research shows that subsurface cracks play a dominant role in the formation of spalling failure. These cracks originated from volume defects of the material and propagate,to form elliptical fatigue spalls under the action of principal tensile stresses. The principal tensile stress increases with increasing contact load,causing spall formation and reduction of rolling contact life. The greater the principal tensile stress is,the more severe the peeling of near surface is. Under the same condition,the closer volume defects are to the surface,the more likely failure occurs,the shorter the rolling contact life is.     

Application of vibration signal measurement for monitoring grinding parameters

Vibration signal measurements are discussed for monitoring operating parameters in a laboratory-scale ball mill. The vibration signal was first picked up with an accelerometer, amplified by a vibrometer and then transmitted to a DAT recorder during the entire testing period. The signal on the DAT recorder was resampled and converted into IBM compatible personal computer readable data format using a digital oscilloscope. The vibration signal analyses included rms estimation, power spectral estimation, waterfall plot, principal component analysis and stepwise multiple regression analysis. Clear differences in the rms and the spectra are found for different grinding conditions. Three principal components described about 95% of the total variation in the spectra. Each principal component was related mainly to one to three major frequency bands. Close correlation was found between the vibration signal and grinding parameters. Therefore, an alternative method can be developed for monitoring the operating parameters in a ball mill.     

空间应用干涉成像光谱仪的研究

干涉成像光谱技术是空间目标探测的一项新技术,近年来发展迅速。按调制方式干涉成像光谱仪可以分为时间调制、空间调制和联合调制三种。介绍各类干涉成像光谱仪的原理及其发展,并给出了几种典型的干涉成像光谱仪,尤其是LASIS相机,近年来成为干涉成像光谱仪研究的新热点。     

Multi-objective optimization of cutting parameters in high-speed milling based on grey relational analysis coupled with principal component analysis

This paper investigates optimization problem of the cutting parameters in high-speed milling on NAK80 mold steel. An experiment based on the technology of Taguchi is performed. The objective is to establish a correlation among spindle speed, feed per tooth and depth of cut to the three directions of cutting force in the milling process. In this study, the optimum cutting parameters are obtained by the grey relational analysis. Moreover, the principal component analysis is applied to evaluate the weights so that their relative significance can be described properly and objectively. The results of experiments show that grey relational analysis coupled with principal component analysis can effectively acquire the optimal combination of cutting parameters and the proposed approach can be a useful tool to reduce the cutting force.     

LAMOST平衡系统的传感器优化布置

针对模态测试中的传感器优化布置问题,提出一种新的主元分析和组合MAC联合算法。首先,将多种传感器布置方法的结果形成原始特征数据矩阵,利用主元分析获得空间测点的主元和综合评价值,据此可以合理地大幅减少候选测点数目;其次,使用组合MAC法计算所有组合的MAC矩阵,取最大非对角元最小的一组作为最终选择;然后,以悬臂梁为例,分别用联合算法和4种传统算法对其进行传感器优化布置,通过3个评价准则的数值比较,证明联合算法的优越性;最后,将联合算法应用在望远镜LAMOST的平衡系统上,得到了两个方向上的传感器布置方案,进一步验证了联合算法的可行性。     

Optimization of correlated multiple responses of ultrasonic machining (USM) process

Ultrasonic machining (USM) process has several important performance measures (responses), some of which are correlated. For example, material removal rate and tool wear rate are highly correlated. Although in the recent past several methods have been proposed in the literature to resolve the multi-response optimization problems, only a few of them take care of the possible correlation between the responses. All these methods primarily make use of principal component analysis (PCA) to consider the possible correlation between the responses. Process engineers may face the difficulty of selecting the appropriate method because the relative optimization performances of these methods are unknown. In this paper, two sets of past experimental data on USM process are analysed using three methods dealing with the multiple correlated responses, and the optimization performances of these three methods are subsequently compared. It is observed that both the weighted principal component (WPC) and PCA-based TOPSIS methods result in a better optimization performance than the PCA-based grey relational analysis method. However, the WPC method is preferable because of its simpler computational procedure.     

二次主成分提取高光谱的病害薯叶特征波长

针对马铃薯晚疫病,提出了将病害叶片和健康叶片联合测试并提取有效特征波长的检测方法。对健康和病害叶片的光谱图像进行主成分分析,并从主成分图像的权重系数曲线中提取出6个健康叶片特征波长和病害叶片特征波长。基于健康叶片与病害叶片的6个特征波长做第二次主成分分析,将特征波长优化至712.19 nm、749.70 nm和841.47 nm,再基于这3个特征波长做主成分分析,选用主成分中对比度最明显的图像来识别病害区域。研究表明,采用健康叶片与病害叶片联合使用并结合二次主成分分析可实现马铃薯叶片病害的设别,且识别率可达100%。     

Three-dimensional molecular shape determination from a limited number of projections

A real space method allowing the reconstruction of negatively stained crystalline objects from a limited number of projections is presented. The method is based on the assumption that only two density levels are required to describe an ideally negatively stained object (that of the volume occupied by the object and that of the volume occupied by the stain). The method is illustrated by the reconstruction of the asymmetric unit of catalase microcrystals using only the three principal projections. It is shown that other projections can then be predicted to a good approximation.     

Structural and ultrastructural characterization of zebu (Bos indicus) spermatozoa.

The ultrastructure of normal, ejaculated spermatozoa of Bos indicus was studied by means of electron microscopy, being evaluated in two principal parts, the head and the tail. The head is flat, oval or paddle-shaped with a square base, which provides a concave recess for the insertion of the tail. The acrosome tightly covers the anterior two thirds of the nucleus. A distinct unilateral acrosomal bulge was observed along the apical edge of the head. The equatorial region demarcates the acrosome from the post-equatorial region that covers the caudal one third of the nucleus. The classical 9+9+2 fiber pattern which composes the axoneme was observed along three segments of the tail, namely middle, principal and terminal pieces. The axoneme is anteriorly bound by the mitochondrial helix (middle piece) and posteriorly by the fibrous helix (principal piece), except at the terminal piece. The border between the middle piece and principal piece was well defined due to the termination of the thick mitochondrial helix and the presence of the annulus. Some of the spermatozoa presented cytoplasmatic droplets, which appeared as stalk-like appendages.     

核函数主元分析及其在齿轮故障诊断中的应用

提出了基于核函数主元分析的齿轮故障诊断方法。该方法通过计算齿轮振动信号原始特征空间的内积核函数来实现原始特征空间到高维特征空间的非线性映射。通过对高维特征数据作主元分析,得到原始特征的非线性主元,以所选的非线性主元作为特征子空间对齿轮工作状态进行分类识别。用齿轮在正常状态、裂纹状态和断齿状态下的试验数据对该方法进行了检验,比较了主元分析与核函数主元分析的分类效果。结果表明,核函数主元分析能有效的检测裂纹故障的出现,正确区分不同的故障模式,更适于提取故障信号的非线性特征。     

欠秩并联机构运动螺旋系主螺旋的虚拟机构识别

研究了一种常用的欠秩３—ＲＰＳ并联平台机构。提出采用虚拟机构的方法建立了雅可比矩阵, 又采用二次曲线分解理论识别了三阶螺旋系的主螺旋,讨论了３—ＲＰＳ机构在两种位形下的输出运动螺旋螺距ｈ 的各种取值,这对于从全局上了解机构的运动特性、轨迹规划和控制具有重要意义     

扩张调节片在多种载荷条件下的数值分析

战斗机发动机喷管技术是决定它良好的隐身性、高机动性、敏捷性等性能的主要因素。高推重比的要求使集收扩喷管普遍采用薄璧结构。这种结构在高温、高压作用下不但容易强度失效 ,而且还会导致结构变形过大。喷管的主要构件之一——扩张调节片的变形会大大降低战斗机的性能 ,甚至导致事故。本文建立了扩张调节片有限元模型 ,用有限元法计算出该构件的温度场 ,并对其进行不同载荷条件下 (分布载荷、热载荷、分布载荷和热载荷联合作用 )的应力和热变形分析。结果表明 :热载荷是影响该构件强度和刚度的主要因素 ;该构件强度符合要求 ,刚度有待提高。最后 ,提出了一种考虑热变形等因素影响的运动副间隙确定方法     

机械故障信号的分离

针对机械故障信号经常是多种故障信号的混合，给正确的故障识别造成很大困难的实际情况，提出基于神经网络非线性主分量分析的机构故障信号分离方法。阐述了故障信息的分离与主分量分析的关系。并将二者统一起来，从理论上证明应用主分量分析方法进行故障分离的有效性；介绍神经网络非线性分离，取得令人满意的结果。     

Nonlinear bending vibrations of a rotating rod

The vibrations of a rod with a circular cross section revolving around its axis with immovable supports in the longitudinal direction are examined. A harmonic load orthogonal to the longitudinal axis and located in a plane rotating jointly with it acts on the rod. The geometrical nonlinearity caused by medial line length variation is taken into account. For free oscillations, three solutions are found for amplitude-frequency dependences and the corresponding modes. In the neighborhood of the principal resonance, the dependences between oscillation amplitudes in orthogonal planes and the angular velocity of rotation under constant frequency of the load are constructed.     

OPTIMIZATION STUDY IN MANUFACTURING PROCES SFOR PC／ABS BLENDS

The optimization of injection molding process for polycarbonate/acrylonitrile-butadiene-styrene (PC/ABS) blends is studied using Taguchi method and principal component analysis (PCA). Four controllable process factors are studied at three levels each in the manufacturing process. The L9 orthogonal array is conducted to determine the optimum process factor/leve! combination for single quality of mechanical properties. In addition, the principal component analysis is employed to transform the correlated mechanical properties to a set of uncorrelated components and to evaluate a comprehensive index for multi-response cases. Then the optimum process factor/level combination for multiple qualities can be determined. Finally, the analysis of variance is used to find out the most influential injection molding parameter for single and multiple qualities problems.