一种基于典型相关分析的特征融合算法

针对声呐目标识别的特征融合问题,采用一种基于典型相关分析(CCA)的融合算法,实现了不同特征的有效融合,在保留特征有效信息的同时消除了冗余信息。实验数据分析结果表明,该方法及其改进算法一定程度上提高了目标识别正确率。     

应用主成分分析和典型相关分析来分析教学

用主成分分析和典型相关分析的方法，试图对某专业的各门课程之间的相互关联性做一分析，从而对提高教与学的效率和合理安排教学任务提出有益的见解。     

基于核典型相关分析的非线性相关源盲分离方法研究

针对传统的相关源盲分离方法的不足,提出了一种基于核典型相关分析的非线性相关源盲分离方法。该方法是利用了核方法来处理数据之间的非线性问题,同时还利用信号源之间的相关性来进行分离。提出的方法与传统的相关源盲分离方法进行对比分析。仿真结果表明,提出的方法明显优于传统的相关源盲分离方法,并从分离性能指标上得到了充分的反映。最后,将该方法应用到转子不对中和碰摩故障的盲分离中,实验结果进一步验证了该方法的有效性。     

相关分析在特征选择中的应用

目的 研究模式识别中特征选择的理论和方法。方法 利用相关分析法实现高效率的特征选择。 结果 在4种车辆声信号特征选择的实际应用中取得了良好的效果。结论 该方法具有运算量小,所选特征识别效果好等明显优点。     

线性特征值问题在模态坐标系中的矩阵摄动法

本文在文献[1,2]的基础上进一步探讨了对线性广义特征值问题Ax=λBx 进行快速近似重分析的矩阵摄动方法和理论,特别对重特征值情况也完整地导出了特征值和特征向量的直到二阶摄动量的渐近估计算式,并在E.J.Haug 等证明了重特征值的方向可微性的基础上指出对应于重特征值的特征向量在一定意义上也是方向可微的。算例表明本文的算法是正确、有效的。     

非保守系统特征灵敏度分析直接法

当结构动力学系统的阻尼矩阵不能同时通过质量和刚度矩阵对角化时,线性振动系统的特征值问题就转化为二次特征值方程,相应的特征值和特征值向量以及它们的导数都成为复空间内的量.针对非保守系统的二次特征值问题,通过求解非齐次线性方程组,直接导出非保守系统特征值和特征向量的一阶灵敏度.提出的方法不需要非保守系统的全部模态,因此,适用于大型复杂结构的特征灵敏度分析.     

多彩色空间相关分析的人脸识别算法

人脸识别是当前人工智能和模式识别的研究热点,得到了广泛的关注.基于对不同色彩空间数据的分析,论文提出了多彩色空间典型相关分析的人脸识别方法.文中对2维的Contourlet变换特性进行了分析和讨论,利用Contourlet的多尺度,方向性和各向异性等特点,提出了一种基于Contourlet变换的彩色人脸识别算法.算法对原图进行Contourlet分解,对分解得到的低频和高频图像进行cca分析.典型相关分析是一种有效的分析方法,其实际应用十分广泛.低频系数反映图像的轮廓信息,高频系数反映图像的细节信息,使用cca充分利用不同频率的信息,使不同色彩空间的不同分辨率图形的相关性达到最大,得到投影系数,最后,采用决策级最近邻分类器完成人脸识别.在对彩色人脸数据库AR的识别实验中,该算法识别率达到98％以上,与传统算法相比,该算法不仅既有良好的识别结果,而且具有很快的运算速度.     

基于典型谱相关峭度图的滚动轴承故障诊断方法EI北大核心CSCD

针对工程实际中滚动轴承发生故障的类型具有典型性和故障冲击信号具有周期性的特点,提出了一种典型谱相关峭度图算法。该算法在借鉴典型谱峭度图算法区间划分的思想基础上,将相关峭度指标代替峭度指标,不但避免了宽频带解调引入的噪声干扰,而且充分利用了典型故障冲击的周期性信息,并通过优化谱相关峭度值,快速定位典型故障冲击信号所在的频率区间,并将该算法应用于最优解调频带的确定。通过对仿真信号和轮对轴承实验信号的分析表明,该算法无论在准确性还是在稳定性方面均表现出了极大的优越性,能够有效的自适应定位共振频带。     

我国深部典型煤与瓦斯突出事故分析

随着煤矿开采深度的增加,煤炭开采过程中出现了一些新现象,特别是煤与瓦斯突出的现象日趋增多。本文对我国发生的几起典型的煤与瓦斯突出事故进行分析,提出煤与瓦斯事故的一些典型特征,以期为煤矿安全生产提供参考。     

建筑施工现场安全管理相关问题分析

近年来,建筑工程的建设规模越来越大,其建设质量关系到建筑行业的健康发展,也与社会经济以及人们的生命财产安全具有十分密切的联系。但是建筑行业也属于一项高危性行业,每年因建筑工程施工死亡的人数一直居高不下。究其主要原因是由于建筑工程施工现场安全管理方面出现了问题。因此我们在实际工作中必须要对施工现场的安全管理引起高度重视。本文就建筑施工现场安全管理相关问题进行分析,以供参考。     

On invariance in canonical analysis with applications to covariate discriminant analysis

We give a result about the invariance of canonical analysis of Euclidean subspaces, so improving a known property. Furthermore, we apply this result to determine conditions for invariance of covariate discriminant analysis when the original variables are linearly transformed. That permits us to introduce a test for this invariance for which we only suppose that the variable admits fourth order moments. A test for redundancy of subcomponents’ variables is obtained as a particular case.      

Two-dimensional correlation analysis

This tutorial reviews the recent computational advances in two-dimensional (2D) correlation spectroscopy, presents the theory, and provides examples applying 2D correlation analysis. Two-dimensional correlation analysis is a method for visualizing the relationships among the variables in multivariate data and their temporal behavior by applying the complex cross-correlation function. This function measures correlations that occur at the same rate or frequency with respect to the data acquisition time. The complex cross-correlation function yields real and imaginary components that contain information about the phase behavior of the variables. The real component provides information about mutually dependent in-phase variations. Variations that occur out-of-phase (with time lags or leads) are given by the imaginary component. Two-dimensional correlation analysis is a general analysis method that can be used for the treatment of data from a variety of applications including image, distribution, environmental, and kinetic analysis.     

Spectrally resolved fluorescence correlation spectroscopy based on global analysis

Multicolor fluorescence correlation spectroscopy has been recently developed to study chemical interactions of multiple chemical species labeled with spectrally distinct fluorophores. In the presence of spectral overlap, there exists a lower detectability limit for reaction products with multicolor fluorophores. In addition, the ability to separate bound product from reactants allows thermodynamic properties such as dissociation constants to be measured for chemical reactions. In this report, we utilize a spectrally resolved two-photon microscope with single-photon counting sensitivity to acquire spectral and temporal information from multiple chemical species. Further, we have developed a global fitting analysis algorithm that simultaneously analyzes all distinct auto- and cross-correlation functions from 15 independent spectral channels. We have demonstrated that the global analysis approach allows the concentration and diffusion coefficients of fluorescent particles to be resolved despite the presence of overlapping emission spectra.     

谱相关密度分析在轴承点蚀故障诊断中的研究

利用谱相关密度提取轴承故障特征时需要在循环频率域和频率域上同时兼顾高分析带宽和高分辨率,从而使得该方法的计算量庞大,难以达到较高的分析精度.鉴于此,首次在循环平稳分析中引入解析的思想,利用解析形式的谱相关密度在循环频率域不存在高频特征的特点,提出运用时域选抽技术,在保证分辨率的同时降低分析带宽,减少计算量,从而得到更好的分析效果.本文以一般调幅信号解析形式的谱相关密度分析为基础,对滚动轴承点蚀故障模型进行了分析,推导了其谱相关密度分析的理论结果,给出具体的算法实现.仿真调幅模型和实际轴承故障信号,证实了理论分析的正确性和算法的可行性,同时也验证了谱相关密度分析对调幅特征的提取能力.     

相关Hilbert差值时延估计方法误差分析

被动探测的关键技术之一就是通过有效的信号处理方法提取出多接收器的时间差,而确定时差最基本的方法是进行信号相关以获得估计信息,相关算法虽然计算过程简单,但是算法对信号和噪声频谱特性要求比较高。基于希尔伯特变换的时延估计是广义相关时延估计方法的拓展,把相关函数的对称性由偶函数转化成奇函数,将相关法中对峰值的检测转化为对过零点的检测。通过对两种算法的误差比较分析,信号的截至频率越高,希尔伯特变换法的均方误差与基本相关法的均方误差相差越明显,估计误差越理想,证明Hilbert算法的优越性。     

Fractional correlation operation: performance analysis

The fractional correlation is a new tool related to the fractional Fourier transform. It is useful for comparison and recognition, especially for shift-variant cases. The performances of such a correlator are analyzed according to the standard criteria of signal-to-noise ratio, correlation sharpness (peak-to-correlation energy), and Horner efficiency. The conclusions are that the performance is object dependent and that for nonwhite noise, compared with the conventional correlator, improved performances are possible. In addition we show that for a white-noise spectrum the fractional correlation has performances similar to the conventional correlator.     

一种基于灰关联分析的语音/音乐分类方法

针对音频信号的特点,提出了一种将灰关联分析应用于语音/音乐信号的分类方法,并给出了对音频信号进行灰关联分析的方法和步骤。利用语音和音乐信号的短时能量均方根的概率统计特征值建立了目标的参考数据和比较数据,进行了不同语音和音乐信号的灰关联分析,确定了目标分类的判据,并对两类信号以及含有独立两类信号的音频信号进行了分类。仿真结果表明:基于单特征值的音频信号灰关联分析方法实现过程简单,分类平均准确率达到90%,优于文献[4]中基于1种和2种特征值的分类性能。     

A hybrid classification method: discrete canonical variate analysis using a genetic algorithm

This paper describes a novel, hybrid multivariate classification method: discrete canonical variate analysis (DCVA), which is integrated in the present implementation with a genetic algorithm (GA). DCVA transforms a multivariate data set into a set of discrete scores of lower dimensionality, intended specifically to act as classifiers of observations into one out of multiple pre-defined groups. The condition for selecting the DCVA loadings is maximization of the ratio of the between-groups to within-groups variance of the scores, but unlike conventional CVA, there is a non-linear, discontinuous relationship between the scores and loadings. The performance of the DCVA method is compared with that of two competing classification methods, Artificial Neural Networks (ANNs) and Mahalanobis distance-based Linear discriminant analysis (LDA) using six example problems. In all cases, internal (leave-one-out) cross-validation was used, and classification success rates retained from both the training and test segments. Of the methods studied, DCVA clearly performed the best in training, producing the highest mean success rates for four out of the six example data sets. For the test segments, DCVA produced the best performance for two of the data sets, and equalled that of LDA and ANN for a third. However, LDA produced the best performance from the remaining three data sets. This is suggestive of a greater tendency of DCVA, like other search-based methods, to overfit.     

线性正则变换的离散化研究进展

线性正则变换(LCT)是Fourier变换和分数阶Fourier变换的广义形式。近年来研究成果表明, LCT在光学、信号处理及应用数学等领域有广泛的应用, 而离散化成为了其得以应用的关键。由于LCT的离散算法不能简单直接地将时域变量和LCT域变量离散化得到, 因此LCT的离散算法成为近年来的研究重点。本文依据LCT的离散化发展历史, 对其重要研究进展和现状进行了系统归纳和简要评述, 并给出不同离散化算法之间的区别和联系, 指明了未来发展方向。这对研究者全面了解LCT离散化方法具有很好的参考价值, 可以进一步促进其工程应用。

     

Quantitative simulation of errors in correlation analysis

Bounding the errors of measurements derived from correlation functions of light scattered from some physical systems is typically complicated by the ill conditioning of the data inversion. Parameter values are estimated from fitting well-chosen models to measurements taken for long enough to look acceptable, or at least to yield convergence to some reasonable result. We show some simple numerical simulations that indicate the possibility of substantial and unanticipated errors even in comparatively simple experiments. We further show quantitative evidence for the effectiveness of a number of ad hoc aspects of the art of performing good light-scattering experiments and recovering useful measurements from them. Separating data-inversion properties from experimental inconsistencies may lead to a better understanding and better bounding of some errors, giving new ways to improve overall experimental accuracy.