基于小波变换、二维主元分析和EBF神经网络的指纹识别方法

指纹识别是模式识别中的一个十分重要课题。结合小波变换(WT)、二维主元分析(2DPCA)和椭球基函数（EBF）特点,本文提出了一种基于WT和2DPCA的EBF神经网络指纹识别方法。首先,利用小波变换将原始图像分解为高频分量和低频分量,并忽略水平高频与垂直高频分量,获得原始图像的基本特征。然后,通过2DPCA算法对该图像进行降维,获取降维特征;最后结合椭球基函数神经网络(Ellipsoidal Basis Function Neural Network, EBFNN)完成指纹识别。本算法将2DPCA优化的特征提取与EBFNN的自适应性相结合,在FVC2000（国际指纹竞赛数据库）上作了测试。并与WT-PNN算法和WT-2DPCA-RBF算法进行比较。实验结果表明,本文提出的算法在平移、旋转及光照变化的指纹数据库上的识别效果优于WT－PNN算法和WT-2DPCA-RBF算法。     

基于快速三维主成分分析的肺CT图像检测

为解决基于二维图像处理的计算机辅助诊疗系统(CAD)仅考虑每幅图像自身包含的信息而忽略不同扫描层之间的联系,以及数据处理过程中的海量计算问题,提出一种新的基于快速三维主成分分析(3D PCA)的有效肺CT病灶检测算法。该算法首先引入高维张量奇异值分解(HOSVD)设计3D PCA;然后以提取出的三维空间特征点为种子点,进行区域增长以获取完整的疑似病灶区域;最后,根据医学图像具体特征,设计了一种HOSVD的简化分解算法。对来自吉林省肿瘤医院的10个典型病例的五百余幅临床CT图像进行了实验,并将实验结果与当前同类算法做了比较。结果表明,检测精确度提高了约10%~21%;另外,快速算法与原算法比较,计算复杂度可降低约1/3。由于快速3D PCA可以挖掘更多存在于不同连续扫描层间的有用信息,更精准提取病灶特征,在一定程度上提高了检测率。     

Automated inspection of PCB components using a genetic algorithm template-matching approach

Automated inspection of surface mount PCB boards is a requirement to assure quality and to reduce manufacturing scrap costs and rework. This paper investigates methodologies for locating and identifying multiple objects in images used for surface mount device inspection. One of the main challenges for surface mount device inspection is component placement inspection. Component placement errors such as missing, misaligned or incorrectly rotated components are a major cause of defects and need to be detected before and after the solder reflow process. This paper focuses on automated object-recognition techniques for locating multiple objects using grey-model fitting for producing a generalised template for a set of components. The work uses the normalised cross correlation (NCC) template-matching approach and examines a method for constraining the search space to reduce computational calculations. The search for template positions has been performed exhaustively and by using a genetic algorithm. Experimental results using a typical PCB image are reported.     

Three-dimensional metrology of surface extracted from a cloud of measured points using a new point-to-surface assignment method: An application to PCB-mounted solder pastes

This work presents a method of point-to-surface assignment for three-dimensional metrology of solder paste bricks on printed circuit board (PCB). A bounding box enclosing the brick tightly on all sides is introduced to avoid incorrect point-to-surface assignment. The shape of bounding box for solder paste brick is variable according to geometry of measured points. The surface geometry of the bounding box is obtained by using five peaks selected from the histogram of normalized gradient vectors for measured points. By using the bounding box enclosing the solder paste brick, the task of point-to-surface assignment is successfully executed. Subsequently, the geometrical features are obtained via surface fitting.     

Support vector regression for determining the minimum zone sphericity

Several methods have been investigated to determine the deviation of manufactured spherical parts from ideal geometry. One of the most popular is the least squares technique, which is still widely employed in coordinate measuring machines used by industries. The least squares algorithm is optimal under the assumption that the data set is very large and has the inherent disadvantage of overestimating the minimum tolerance zone, resulting sometimes in the rejection of good parts. In addition, it requires that the data be distributed normally. The support vector regression approach alleviates the necessity for these assumptions. While most fitting algorithms in practice today require that the sampled data accurately represent the surface being inspected, support vector regression provides a generalization over the surface. We describe how the concepts of support vector regression can be applied to the determination of tolerance zones of nonlinear surfaces; to demonstrate the unique potential of support vector machine algorithms in the area of coordinate metrology. In specific, we address part quality inspection of spherical geometries.     

Fast and accurate NURBS fitting for reverse engineering

Fast and accurate fitting of non-uniform rational B-spline (NURBS) curves and surfaces through large sets of measured data is an important problem in applications such as reverse engineering and geometric modelling. This paper presents a method for realising significant improvements in the computational efficiency of this task. The basic idea is that the sparsity structures of the relevant matrices that are specific to the problem of NURBS fitting can be precisely defined and that full exploitation of these structures leads to significant savings in both computational and storage requirements. These savings allow for a large number of control points to be used in order to define the surface and consequently to improve the accuracy of shape representation. The achieved computational complexity is linear in both the number of measured points and the number of control points while the storage requirements of the algorithm are linear with the number of control points only. The complexity analysis, as well as the analysis of actual running times is presented. The results demonstrate that, using this approach, highly complex shapes may be modelled accurately with a single NURBS surface.     

Circle fitting from the polarity transformation regression

Geometrical fitting is useful in different fields of science and technology, in particular least squares minimum (LSM) methods are widespread in contact probing for coordinate measuring machines, as well as a reference shape for surface metrology. We present a new intuitive and simple LSM algorithm for circle fitting, the polarity transformation regression. It is a non-linear algebraic method from a generic geometric transformation. We derive the explicit expression of the model estimators from the data points. Then, the algorithm is compared with other methods based on simulation and some literature data sets. The proposed algorithm presents a comparable accuracy, low computational effort and good behavior with outliers based on the initial test, outperforming other well-known algebraic methods in some of the studied data sets. The basis of the algorithm is finally suggested for other potential uses.     

基于支持向量机的点云曲面刀具路径规划

针对单值散乱点云曲面刀具路径规划问题,提出了一种基于最小二乘支持向量机的计算方法。在计算过程中,将点云数据向平面投影,得到二维点集。应用网格划分和边界网格内测量点高斯映射技术,提取平面区域内的边界特征点。用边界特征点定义点云曲面的实际加工区域,在此区域内规划平行等间距刀具路径。应用最小二乘支持向量机拟合点云数据,求得被加工曲面的连续表达模型,经此模型将二维刀具路径数据向三维空间映射,求出刀触点数据。将刀触点经法向偏置计算,求得刀位点。实例验证证明,该方法能较好地解决信息不完备散乱点云曲面刀具路径生成问题。      

基于遗传算法的圆度公差评定法与采用最小二乘法评定的比较

根据提出的计算模型,对基于遗传算法的圆度误差评定和传统上采用最小二乘法的评定算法进行了比较分析,根据方法本身的特点和计算结果,分析了二者的不同点以及在工程应用中的适用场合.所构造的模型包括边界控制点和区域随机点,其中边界控制点模拟了由圆度误差最小区域条件所定义的最大内切圆和最小外切圆,而区域随机点模拟了实际情况下测试点的随机性和不确定性.计算结果表明基于遗传算法的圆度评定法精度较高,优于基于最小二乘法的评定算法.     

Method for cylindricity error evaluation using Geometry Optimization Searching Algorithm

According to the geometrical characteristics of cylindricity error, a method for cylindricity error evaluation using Geometry Optimization Searching Algorithm (GOSA) has been presented. The optimization method and linearization method and uniform sampling could not adopt in the algorithm. The principle of the algorithm is that a hexagon are collocated based on the reference points in the starting and the end measured section respectively, the radius value of all the measured points are calculated by the line between the vertexes of the hexagon in the starting and the end measured section as the ideal axes, the cylindricity error value of corresponding evaluation method (include minimum zone cylinder method (MZC), minimum circumscribed cylinder method (MCC) and maximum inscribed cylinder method (MIC)) are obtained according to compare, judgment and arranged hexagon repeatedly. The principle and step of using the algorithm to solve the cylindricity error is detailed described and the mathematical formula and program flowchart are given. The experimental results show that the cylindricity error can be evaluated effectively and exactly using this algorithm.     

Multi-variable measurements and optimization of GMAW parameters for API-X42 steel alloy using a hybrid BPNN–PSO approach

This research addresses multi criteria modeling and optimization procedure for Gas Metal Arc Welding (GMAW) process of API-X42 alloy. Experimental data needed for modeling are gathered as per L₃₆ Taguchi matrix. Model inputs include work piece groove angle as well as the five main GMAW process parameters. The proposed back propagation neural network (BPNN) simultaneously predicts weld bead geometry (WBG) and heat affected zone (HAZ). Image processing technique along with Bridge Cam and AWS gauges are used to take accurate measurements of WBGs and HAZs. The adequacy of the developed BPNN is established through comparisons against measured process outputs. Measurements indicate that the BPNN model simulates GMAW process with average errors of 0.33 to 0.82%. Next, the BPNN model is implanted into a particle swarm optimization (PSO) algorithm to simultaneously optimize HAZ and WBG characteristics. The hybrid BPNN–PSO determines process parameters values and groove angle so as a desired WBG is achieved while HAZ is minimized. Verification tests demonstrate that the proposed BPNN–PSO is quite efficient for in multi-criteria modeling and optimization of GMAW.     

焊接缺陷的磁光成像小波多尺度识别及分类

针对焊缝微小凹陷、未熔合和焊偏等焊接缺陷,提出了基于磁光成像无损探伤的小波多尺度边缘提取算法及主成分分析-误差反向传播神经网络(PCA-BP)缺陷分类模型;研究了焊件表面及近表面缺陷的可视化无损检测及分类方法。首先,通过对焊件施加感应磁场,利用法拉第磁致旋光原理构成磁光传感器,获取焊接缺陷磁光图像。然后,针对焊接缺陷磁光图像存在噪声干扰、对比度低且成像背景复杂等特征,基于小波模极大值的多尺度边缘信息融合方法,设计了具有高抗噪性的缺陷边缘检测算法。最后,通过PCA法对磁光图像列方向灰度变量进行预处理,得到能表征95%磁光图像列方向灰度变量信息的256个特征点作为输入特征量,构建了三层BP神经网络模型,对焊接缺陷样本进行分类。试验结果表明,所提方法能准确识别微小凹陷、未熔合和焊偏等焊接缺陷,模型分类准确率可达90.80%。     

Reverse engineering of turbine blades based on design intent

The purpose of this paper is to describe a new approach to process the data points measured from turbine blade airfoils in order to make a valid shape via reverse engineering method. Currently, preliminary B-rep models can be created by fitting surfaces to point clouds using a 3D laser scanner. In case of a turbine blade, due to high shape complexity, the resulting model is often unsuitable in practice. A small change in blade geometry can lead to a large change in turbine performance. Therefore, control of the blade shape is critical to the design process. Authors believe that the only way to capture the valid shape of a blade airfoil out of the many manufacturing deviations is to incorporate design key-points during reverse engineering. Implementation of the new method using segmentation and constrained fitting algorithm (SCFA) on a heavy-duty industrial gas turbine blade has been reported and discussed.     

Experimental resolution measurement in critical dimension scanning electron microscope metrology

By applying the basic principles of metrology we discuss how to define the standards that any experimental method to measure resolution has to obey. Our results clearly indicate the need to apply a calibration procedure when designing algorithms to estimate resolution to satisfy accuracy requirements. Similarly, the precision of an algorithm has to be clearly specified. We compare here the performances of a variety of commonly used implementations of published methods, with that of an algorithm based on an approach known to be reliable. Our results confirm that when an algorithm is designed with the clear intent of satisfying metrology requirements, it demonstrates excellent accuracy, precision, and lack of sensitivity to the noise level, as is desirable. As a consequence, the algorithm will have the ability to measure accurately the point spread function convoluted in the image, thus paving the way for quantitative deconvolution techniques.     

Efficient model building in active appearance model for rotated face

This paper proposes the efficient model building in active appearance model （AAM） for the rotated face.Finding an exact region of the face is generally difficult due to different shapes and viewpoints.Unlike many papers about the fitting method of AAM,this paper treats how images are chosen for fitting of the rotated face in modelling process.To solve this problem,databases of facial rotation and expression are selected and models are built using Procrustes method and principal component analysis （PCA）.These models are applied in fitting methods like basic AAM fitting,inverse compositional alignment （ICA）,project-out ICA,normalization ICA,robust normalization inverse compositional algorithm （RNIC） and efficient robust normalization algorithm （ERN）.RNIC and ERN can fit the rotated face in images efficiently.The efficiency of model building is checked using sequence images made by ourselves.     

基于三维凸包的公差基准平面拟合方法

针对目前基于离散采样点进行基准平面拟合时没有考虑基准平面受约束的问题,提出一种基于三维凸包的带约束的基准平面拟合方法。该方法利用计算几何中的凸包理论,使用增量算法快速构建平面特征离散采样点的三维凸包。其次根据基准平面的定义,确定基准平面处于不同基准优先级时被其他基准特征约束的自由度以及建立基准的方法。在此基础上利用三维凸包建立基准平面无约束、转动约束、平动约束的数学模型,结合基准的相关实体要求实现了基准平面变动空间的求解。试验结果表明：采用一组采样数据建立基准平面时,所提出的算法与传统的最小二乘法进行比较更加符合基准平面的建立原则,保证了基准平面的准确性,满足了工程需求。     

A robust surface fitting and reconstruction algorithm for form characterization of ultra-precision freeform surfaces

Ultra-precision freeform surfaces are non-rotational symmetric surfaces possessing sub-micrometer form accuracy and nanometric surface finish. Although they can be fabricated accurately by ultra-precision machining technology, their surface quality is difficult to be characterized. Surface reconstruction is a vital task in the form characterization of ultra-precision freeform surfaces. This paper presents a robust surface fitting algorithm to reconstruct a high fidelity surface from measured discrete points while the surface smoothness can be ensured as well. A fitting threshold named confidence interval of fitting error is proposed to strike the balance between fitting accuracy and surface smoothness in the fitting process. The fitting algorithm is in two steps. In the first step, bidirectional sampling method is developed to extract a curve network from measured points cloud to construct an initial surface. In the second step, the fitting error of the initial surface is minimized to meet the prescribed fitting error threshold. A series of experimental work has been conducted and the results show that the proposed algorithm is able to provide effective means for increasing the accuracy in the form characterization of ultra-precision freeform surfaces.     

A reliable method of minimum zone evaluation of cylindricity and conicity from coordinate measurement data

The form error evaluation of cylinders and cones is very important in precision coordinate metrology. The solution of the traditional least squares technique is prone to over-estimation, as a result unnecessary rejections may be caused. This paper proposes a reliable algorithm to calculate the minimum zone form errors of cylinders and cones, called a hybrid particle swarm optimization-differential evolution algorithm. The optimization is conducted in two stages, so that the program can hold a fast convergence rate, while effectively avoiding local minima. Experimental results demonstrate that the proposed algorithm can obtain very accurate and stable results for the calculation of cylindricity and conicity.     

基于独立分量分析的切削声发射源信号分离

针对切削声发射(Acoustic Emission,AE)信号的多目标状态源并行分离问题和同频干扰源分离问题,引入独立分量分析(Independent Component Analysis,ICA)技术作为研究工具,用刀具破损、切屑折断和环境噪声三个AE源的线性混合模拟切削AE信号,尝试用FastICA算法分离目标状态...     

Conicity error evaluation using sequential quadratic programming algorithm

Form error evaluation plays an important role in processing quality evaluation. Conicity error is evaluated as a typical example in this paper based on sequential quadratic programming (SQP) algorithm. The evaluation is carried out in three stages. Signed distance function from the measured points to conical surface is defined and the cone is located roughly by the method of traditional least-squares (LS) firstly; the fitted cone and the measured point coordinates are transformed to simplify the optimal mathematical model of conicity error evaluation secondly; and then optimization problem on conicity error evaluation satisfying the minimum zone criterion is solved by means of SQP algorithm and kinematic geometry, where approximate linear differential movement model of signed distance function is deduced in order to reduce the computational complexity. Experimental results show that the conicity error evaluation algorithm is more accurate, and has good robustness and high efficiency. The obtained conicity error is effective.