Research of Adaptive Neighborhood Incremental Principal Component Analysis and Locality Preserving Projection Manifold Learning Algorithm

In view of the incremental learning problem of manifold learning algorithm, an adaptive neighborhood incremental principal component analysis (PCA) and locality preserving projection (LPP) manifold learning algorithm is presented, and the incremental learning principle of algorithm is introduced. For incremental sample data, the adjacency and covariance matrices are incrementally updated by the existing samples; then the dimensionality reduction results of the incremental samples are estimated by the dimensionality reduction results of the existing samples; finally, the dimensionality reduction results of the incremental and existing samples are updated by subspace iteration method. The adaptive neighborhood incremental PCA-LPP manifold learning algorithm is applied to processing of gearbox fault signals. The dimensionality reduction results by incremental learning have very small error, compared with those by batch learning. Spatial aggregation of the incremental samples is basically stable, and fault identification rate is increased.     

流形学习算法概述

流形学习是一种非监督学习算法,流形学习算法的目的是挖掘嵌入在高维数据空间中的低维光滑流形,本文在论述流形学习算法诞生及研究现状的基础之上,指明流形学习算法的研究重点:流形本征维数估计、有监督学习、样本外学习能力、特殊流形降维,并指出流形学习的研究意义。     

有限快拍下的小尺度矢量阵DOA算法

针对小尺度矢量水听器阵条件下ESPRIT算法目标方位估计精度低、多目标分辨能力弱的缺点,提出了一种基于酉变换的DOA改进算法.该方法无需计算观测数据的协方差矩阵,利用阵列的观测数据及共轭信息构造一中心Hermitian对称矩阵,根据酉变换原理将其转化为实数矩阵并用实值ESPRIT方法估计目标方位.由于利用了观测数据的共轭信息,相当于虚拟增加了阵元个数,改进算法有更高的DOA估计精度和多目标分辨能力,在快拍数与阵元数相当时改进算法比协方差方法有更小的计算复杂度.仿真分析结果证明了算法的有效性.     

基于小波变换和线性判别分析的人脸识别方法

提出了将人脸图像的小波分解和线性判别分析结合以达到人脸识别的方法.首先对人脸图像作小波分解,并将分解后的低频系数进行线性判别分析进一步降低人脸特征向量的维数,最后利用最近邻分类器进行分类识别.实验表明,该方法的正确识别率高于传统的特征脸识别方法.     

Unusual event detection and prediction in real-life scenes

In this paper,we consider unusual event detection problem in a novel viewpoint and provide an algorithm to solve the problem.The actions or events in the scene is usual or not will eventually be reflected on the changes of some basic features.We summarize these basic event features and propose special representation for each of them.Thus we can model these features in a uniform mode using adaptive Gaussian mixture model.Supervised and unsupervised unusual event detection algorithm can be designed to fit var...     

Unsupervised color-texture image segmentation

The measure J in J value segmentation （JSEG） fails to represent the discontinuity of color, which degrades the robustness and discrimination of JSEG. An improved approach for JSEG algorithm was proposed for unsupervised color-texture image segmentation. The texture and photometric invariant edge information were combined, which results in a discriminative measure for color-texture homogeneity. Based on the image whose pixel values are values of the new measure, region growing-merging algorithm used in JSEG was then employed to segment the image. Finally, experiments on a variety of real color images demonstrate performance improvement due to the proposed method.     

一种新型多核判别分析方法

为了给模式分类和维度约简提供有效的手段,在综合L1-MKDA和L2-MKDA两种多核KDA优点的基础上,提出一种以预定内核函数的线性组合,并结合混合范数正则化函数实现核权重的稀疏性和非稀疏性平衡的新型弹性多核判别分析方法(semi-infinite-programming-based flexible multi-kernel discriminant analysis, S-FMKDA).该方法用半无限规划方法求解弹性多核判别分析算法,并通过混合正则化实现核的自学习.在不同数据集上的实验结果表明:S-FMKDA比目前常见的KDA、KDAP、KDAG、L1-MKDA、L2-MKDA、UMKDA核判别分析方法的精度提高5%.      

一种由RBF网络驱动的模糊产生式系统

介绍了ＲＢＦ网络和模糊产生系统，分析了由ＲＢＦ网络实现模糊产生式系统的条件，并研究了ＲＢＦ网络的训练算法，以确定模糊系统的参数，运用这种学习算法，可同时确定模糊系统规则的数目和规则的参数。仿真结果说明了该方法的有效性。     

遥感在缅甸飓风灾害土地覆盖分类中的应用

为探讨"纳尔吉斯(Nargis)"飓风对缅甸伊洛瓦底江三角洲地区的影响,利用该地区的遥感数据,用ISODATA非监督分类和贝叶斯最大似然监督分类方法对研究区土地覆盖进行分类,得到了纳尔吉斯飓风前后缅甸伊洛瓦底江三角洲地区土地覆盖的定量变化,并对纳尔吉斯飓风灾害与红树林面积和密度的关系进行了定量研究.结果表明,红树林面积越大,种植密度越高,对减小飓风灾害的作用越大.     

一种机动目标的自适应跟踪算法

通过分析"当前"统计模型采取固定机动加速度最大值在目标跟踪中的不足,引入模糊逻辑技术,提出了一种基于"当前"统计模型的自适应跟踪算法.该算法实现了目标机动加速度最大值随机动特性的自适应调整,从而实现了系统状态噪声方差的自适应调整,提高了系统的跟踪性能.对两种典型的目标机动形式进行了Monte Carlo仿真研究,结果表明了新算法对于解决机动目标跟踪问题的有效性.     

基于三维模型重构技术的公路预制构件尺寸检验评价方法

为使公路工程异形混凝土预制构件的制造尺寸检验评价适应工业化建造要求，应用三维模型重构技术检验评价了异形混凝土预制构件尺寸，提出了高精度、自动化异形混凝土预制构件尺寸检验评价方法，包括三维模型重构、点云数据处理和检验评价体系3个环节；总结了基于三维摄影的点云模型重构技术的原理和关键环节，研究了基于坐标转换和包围盒的无关点云自动剔除算法，并通过2个实例验证了该算法的效果；研究了点云整体配准和3类局部配准方法，针对不同形状的构件和不同位置的细部构造，结合工程需求进行对照；考虑工程需求并与配准方法对应，提出了基于彩色误差云图和数理统计方法进行构件尺寸检验评价的3种判异原则，包括误差均值和标准差判异原则、误差极值判异原则、综合判异原则；综合分析彩色误差云图和误差分布，利用判异原则对箱型涵洞侧墙和管型涵洞侧墙实例的尺寸进行了检验评价。研究结果表明:3种判异原则分别适合对预制构件进行整体宏观、局部单项和先整体后局部的检验评价；采用三维模型重构技术建立的异形混凝土预制构件点云模型，与实际构件相比其长度和宽度的误差均值约为1.5 mm，三维模型重构技术可以替代人工测量；能够自动化剔除无关点云，并执行更为严格且符合工程需求的构件尺寸检验评价。      

一种改进的GPS模糊度白化滤波算法

在深入分析白化滤波去相关算法的基础上,对白化滤波算法进行了改进,以改善GPS模糊度解算中的去相关效果.在改进算法中,对模糊度协方筹矩阵的元素进行排序处理,以保证矩阵分解的有效性,改善去相关效果,提高模糊度离散搜索效率和质量.数值模拟结果表明,该算法不受维数限制,对低维和高维都能成功解算,具有较强的适用性和良好的去相关效果,能提高动态模糊度解算的效率.     

A Fast Algorithm for Support Vector Clustering

Support Vector Clustering (SVC) is a kernel-based unsupervised learning clustering method. The main drawback of SVC is its high computational complexity in getting the adjacency matrix describing the connectivity for each pairs of points. Based on the proximity graph model, the Euclidean distance in Hilbert space is calculated using a Gaussian kernel, which is the right criterion to generate a minimum spanning tree using Kruskal‘s algorithm. Then the connectivity estimation is lowered by only checking the linkages between the edges that construct the main stem of the MST ( Minimum Spanning Tree), in which the non-compatibility degree is originally defined to support the edge selection during linkage estimations. This new approach is experimentally analyzed.The results show that the revised algorithm has a better performance than the proximity graph model with faster speed, optimized clustering quality and strong ability to noise suppression, which makes SVC scalable to large data sets.     

基于无监督学习的部分-整体关系获取

针对面向中文自由文本的部分-整体关系抽取问题,提出一种基于无监督学习的方法. 首先提出子模式提取算法,从领域文本集中获取概念对和概念对所在上下文模式,利用概念对和概念对上下文模式建立分布式语义模型;然后采用协同聚类算法将具有相同语义关系的概念对聚合成簇,通过训练L1正则化逻辑回归模型提取簇的特征并得到代表每个簇语义关系的概念对上下文模式;最后根据模式识别表达部分-整体关系的簇,从而获取部分-整体关系概念对. 实验结果表明,该方法取得较好的性能,F度量达到68.97%,优于传统聚类方法（55.77%）和模式匹配方法（61.95%）.      

基于类间和类内方差的快速二维阈值分割法

为了提高二维阈值分割法的处理速度,提出二维类间方差最大法的快速实现方法.首先,将二维最佳阈值（s*,t*）的求解拆分成两个一维最佳阈值s*和t*的求解,并引入类内距离的定义,提出新的最佳阈值判别式.其次,将原二维直方图分成M×M个区域,合并每个区域为一点,并构建新的二维直方图,在其上应用本文改进的阈值判别式D（s*,t*）求解,得到分割阈值所在的区域编号.最后,在该区域内再次使用D（s*,t*）求解得到原始图像的最佳分割阈值.理论分析及针对不同信噪比的多幅图像的实验结果表明,本文方法的分割错误率低于原始二维Otsu法,且将原算法的时间复杂度由O（L4）降为O（L1/2）,空间复杂度由S（L2）降为S（2L）.      

Iterative algorithm of steered minimum variance and its application in weak targets detection

The steered covariance matrix (STCM) and its inverse matrix should be calculated in each beam for steered minimum variance (STMV). The inverse matrix needs complex computation and restricts its application in engineering. Combining the integration character of one-phase regressive filter with the iterative formula of inverse matrix, an STMV iterative algorithm is proposed. The computational cost of the iterative algorithm is reduced approximately to be 2/M times of the original one when there are M sensors, and is more advantaged for the realization of the algorithm in real time. Simulation results show that the STMV iterative algorithm can preserve the characters of STMV on high azimuth resolution and weak target detection while the computational cost reduced sharply. The analysis on sea trial data proves that the proposed algorithm can estimate each target’s azimuth even when the source powers differ in large scales or their bearings are very approximate.     

基于ICA的MIMO系统空时盲均衡算法

多输入多输出系统的MMSE均衡中,由于接收端采样大小的限制,对观测信号协方差矩阵的估计和信道的盲辨识误差会严重影响均衡性能,采用独立分量分析作为一种辅助手段改进传统的均衡器,并选择最佳均衡时延优化算法。仿真结果表明,该算法提高了传统算法的鲁棒性,在大大降低计算复杂度的同时改善了均衡性能。     

Incremental Supervised Subspace Learning for Face Recognition

Subspace learning algorithms have been well studied in face recognition. Among them, linear discriminant analysis (LDA) is one of the most widely used supervised subspace learning method. Due to the difficulty of designing an incremental solution of the eigen decomposition on the product of matrices, there is little work for computing LDA incrementally. To avoid this limitation, an incremental supervised subspace learning (ISSL) algorithm was proposed, which incrementally learns an adaptive subspace by optimizing the maximum margin criterion (MMC). With the dynamically added face images, ISSL can effectively constrain the computational cost. Feasibility of the new algorithm has been successfully tested on different face data sets.     

基于ICA的MIMO系统空时盲均衡算法

多输入多输出系统的MMSE均衡中,由于接收端采样大小的限制,对观测信号协方差矩阵的估计和信道的盲辨识误差会严重影响均衡性能,采用独立分量分析作为一种辅助手段改进传统的均衡器,并选择最佳均衡时延优化算法。仿真结果表明,该算法提高了传统算法的鲁棒性,在大大降低计算复杂度的同时改善了均衡性能。     

基于改进UKF的BDS/IMU组合列车定位方法

为提高列车定位的精确性和连续性，采用北斗卫星接收机和惯性测量单元构建车载组合定位系统. 针对多传感器组合定位信息融合估计的非线性和鲁棒性需求，将抗差估计理论的等价权原理应用于标准无迹卡尔曼滤波（unscented Kalman filter，UKF）算法，构造了一种改进的UKF算法，通过对标准UKF算法的噪声协方差进行等价替换，从而起到调节滤波增益的作用，使得滤波算法对传感器观测粗差具有较强的抑制能力. 将改进的UKF算法与标准UKF算法应用于列车组合定位进行仿真比较，结果表明:传感器无观测异常时，改进UKF算法的滤波精度总体上略优于标准UKF算法；当传感器观测值含有随机粗差时，改进UKF算法的滤波精度及稳定性明显优于标准UKF算法，北向、东向位置平均估计误差分别降低了48.5%、48.8%，北向、东向速度平均估计误差分别降低了43.7%、48.9%.