基于最大权团的曲面粗匹配算法

提出一种将曲面匹配问题转化为图论中的最大权团搜索问题、将最优的点对应关系用最大权团表示的曲面粗匹配算法,该算法分为点匹配、点对应图构造和最大权团生成等3个阶段.点匹配使用高曲率点和均匀采样点作为候选点,通过自旋图进行匹配计算,构造初始点对应集合;点对应图构造使用距离约束、法矢约束和唯一性约束构造图的边,并使用自旋图相关系数为顶点赋权值;最大权团生成使用基于分支限界的团搜索算法,从对应点图中提取出代表最优对应的最大权团.实验结果表明,文中算法稳定、有效、可扩展,能够进行部分曲面匹配,并且适用于欠特征曲面.     

平面简单闭合曲线离散采样与重建算法

提出一种鲁棒的平面简单闭合曲线离散采样与重建算法。算法分为采样过程和重 建过程两部分。采样部分首先对平面闭合曲线均匀取点,然后计算各点到曲线所围平面区域中 轴的最近距离,最后根据所求距离确定采样间隔,获取采样点集;重建部分首先构建采样点集 的Delaunay 三角剖分,然后从得到的三角形中选择边构建初始化图形,最后通过修改该图形获 得重建图形。实验表明算法得到的采样点较少且能反映曲线的局部几何特性,重建图形能够较 好地表示原闭合曲线的形状及走向。     

拓扑图格独立分量分析和谱聚类支持的纹理探测

提出的纹理探测方法首先采用拓扑图格独立分量分析(TICA)分别对每个观测纹理进行学习,获得分离基．分离基的作用相当于滤波器,并通过最大响应准则得到选择．从不同纹理选择的滤波器构成一个滤波器集;为了计算探测点的纹理特征,测试图像被分解为滤波器通道。最后,探测点被视为谱图的顶点,根据谱聚类(SC)对谱图的切分结果,递归地分离出探测点．实验结果表明,这一方法行之有效。     

Establishing a balanced neighborhood of discrete points for local quadric surface fitting

This paper presents a novel algorithm to establish a balanced neighborhood of points for reliable local quadric surface fitting, a common task in point cloud data processing. The underlying smooth surface geometry of a point cloud in the vicinity of a point can be locally approximated by the best fitted quadric surface at the point. The quality of the fitted surface considerably depends on what neighboring points are selected for the fitting. Specifically, if the selected neighboring points carry a biased distribution, the fitted geometry becomes biased, resulting in loss of accuracy in the fitting. The presented algorithm in this paper is able to reliably select neighboring points considering measures of both distance and direction. The main feature is the development of a geometric relationship, named as Territory Claiming, between the selected and the candidate neighboring points. The fundamental principle is for the selected point set to cover the whole neighborhood domain without redundancy. The selection procedure starts with a distance-based sequence of neighboring points with the territory claiming relationship functioning as a filter to establish a well-balanced neighborhood. The neighborhood can be expanded to incorporate sufficient number of points for the quadric surface fitting while maintaining the balance of the overall neighborhood. The implementation results have demonstrated that the presented method is robust and selects local neighboring points with superior fitting performance in comparison with the distance-based neighbors, mesh neighbors, and elliptic Gabriel graph neighbors.     

Fuzzy points: algebra and application

A fuzzy point is a region representing the uncertain location of a normal Euclidean point. A fuzzy point in the plane is considered to be a closed disk (a circle and its interior). The algebra of fuzzy points (which includes fuzzy vectors and fuzzy angles) is presented. Since fuzzy points are represented as closed disks, the lengths of fuzzy vectors, and the angles between fuzzy vectors can be viewed as properties of circles in the plane. Methods to compute the magnitude of a fuzzy angle are given. An application of fuzzy point algebra to the problem of detecting and tracking storms in Doppler radar image sequences, which motivates this work, is discussed.     

Particle swarm optimization with mutation operations controlled by landscape modality estimation using hill-valley detection

Particle swarm optimization (PSO) is one of swarm intelligence algorithms and has been used to solve various optimization problems. Since the performance of PSO is much affected by the algorithm parameters of PSO, studies on adaptive control of the parameters have been done. Adaptive PSO (APSO) is one of representative studies. Parameters are controlled according to the evolutionary state, where the state is estimated by distance relations among a best search point and other search points. Also, a global Gaussian mutation operation is introduced to escape from local optima. In this study, a new adaptive control based on landscape modality estimation using hill-valley detection is proposed. A proximity graph is created from search points, hills and valleys are detected in the graph, landscape modality of an objective function is identified as unimodal or multimodal. Parameters are adaptively controlled as: parameters for convergence are selected in unimodal landscape and parameters for divergence are selected in multimodal landscape. Also, two mutation operations are introduced according to the modality. In unimodal landscape, a new local mutation operation is applied to the worst hill point which will be moved toward the best point for convergence. In multimodal landscape, a new adaptive global mutation operation is applied to all hill points for escaping from local optima. The advantage of the proposed method is shown by comparing the results of the method with those by PSO with fixed parameters and APSO.     

The relative neighbourhood graph is a part of every 30°-triangulation

We study sets of points in the two-dimensional Euclidean plane. The relative neighbourhood graph (RNG) of a point set is a straight line graph that connects two points from the point set if and only if there is no other point in the set that is closer to both points than they are to each other. A triangulation of a point set is a maximal set of nonintersecting line segments (called edges) with vertices in the point set. We introduce angular restrictions in the triangulations. Using the well-known method of exclusion regions, we show that the relative neighbourhood graph is a part of every triangulation all of the angles of which are greater than or equal to 30°.     

基于多层二部图的高光谱模糊聚类算法

针对传统模糊聚类算法对初始聚类中心非常敏感以及对高光谱图像处理效果不佳的问题,为减少聚类数据的复杂度、降低聚类过程的计算成本以提升聚类性能,提出了一种基于多层二部图的高光谱模糊聚类算法。首先使用SuperPCA预处理方法对超像素分割得到的每个同质区域进行PCA来学习HSI数据不同区域的固有低维特征,从而获得高光谱数据的低维表示;其次,构造一个多层二部图矩阵来描述数据点和锚点之间的关系,降低了计算复杂度;最后,在模糊聚类中加入基于多层二部图的非负正则项来约束模糊隶属度矩阵的解空间。在Indian Pines和Pavia University数据集上进行的实验表明,所提算法能提高聚类效果与性能。     

Delineating tree crowns from airborne laser scanning point cloud data using Delaunay triangulation

Most algorithms for delineating trees from laser scanning data are pixel‐based region growing methods. The algorithm presented in this letter makes use of the original laser points to avoid errors introduced by interpolation. Three different scales are used to identify the seed points or the local maxima at that scale. The seed points are considered to be the estimated tree tops, and are used for growing regions or tree crowns around the seed points. For a test dataset from a Finnish mixed forest and point density of approximately 2 points m^?2, up to 75% of the reference trees could be identified. At the coarser scales, fewer trees were identified, but the crowns were less fragmented. Further work is required to determine how far the method is applicable in other forest conditions and data with other point densities.     

三维图匹配算法

目的 现有的图匹配算法大多应用于二维图像,对三维图像的特征点匹配存在匹配准确率低和计算速度慢等问题。为解决这些问题,本文将分解图匹配算法扩展应用在了三维图像上。方法 首先将需要匹配的两个三维图像的特征点作为图的节点集;再通过Delaunay三角剖分算法,将三维特征点相连,则相连得到的边就作为图的边集,从而建立有向图;然后,根据三维图像的特征点构建相应的三维有向图及其邻接矩阵;再根据有向图中的节点特征和边特征分别构建节点特征相似矩阵和边特征相似矩阵;最后根据这两个特征矩阵将节点匹配问题转化为求极值问题并求解。结果 实验表明,在手工选取特征点的情况下,本文算法对相同三维图像的特征点匹配有97.56%的平均准确率;对不同三维图像特征点匹配有76.39%的平均准确率;在三维图像有旋转的情况下,有90%以上的平均准确率;在特征点部分缺失的情况下,平均匹配准确率也能达到80%。在通过三维尺度不变特征变换（SIFT）算法得到特征点的情况下,本文算法对9个三维模型的特征点的平均匹配准确率为98.78%。结论 本文提出的基于图论的三维图像特征点匹配算法,经实验结果验证,可以取得较好的匹配效果。     

基于图割和模糊连接度的交互式舰船红外图像分割方法

针对舰船红外图像分割中的低对比度、边缘模糊和目标灰度不均匀问题, 提出了基于图割和模糊连接度的交互式图像分割方法. 交互方式为矩形笔刷, 选择目标和背景种子点. 分割方法为基于图割的图像分割方法, 引入模糊连接度来计算图割的似然能, 给出了模糊连接度权重的自动确定方法, 提出了基于直方图分解的高斯混合模型(Gaussian mixture model, GMM)成分个数和参数估计方法. 仿真结果表明, 新方法可实现各种复杂环境下舰船红外图像目标的有效分割.     

K-means初始聚类中心的选择算法

K-means算法随机选取初始聚类中心,容易造成聚类准确率低且聚类结果不稳定。针对这一问题,提出一种初始聚类中心的选择算法。通过k-dist的差值(DK)图分析,确定数据点在k-dist图上的位置,选择主要密度水平曲线上k-dist值最小的点作为初始聚类中心。实验证明,改进算法选择的初始聚类中心唯一,聚类结果稳定,聚类准确率高,迭代次数少。     

Two-dimensional shape decomposition using fuzzy subset theory applied to automated chromosome analysis

This paper describes a technique to transform a two-dimensional shape into a generalized fuzzy binary relation whose clusters represent the meaningful simple parts of the shape. The fuzzy binary relation is defined on the set of convex and concave boundary points, implying a piecewise linear approximation of the boundary, and describes the dissemblance of two vertices to a common cluster. Next some fuzzy subsets are defined over the points which determine the connection between the clusters.The decomposition method first determines nearly convex regions, which are subgraphs of the total graph, and then selects the greatest nearly convex region which satisfies best the defined fuzzy subsets and relations. Using this procedure on touching chromosomes defining the simple parts to be the separated chromosomes, the decomposition often corresponds well to the decomposition that a human might make.     

A linear expected-time algorithm for computing planar relative neighbourhood graphs

A new algorithm for computing the relative neighbourhood graph (RNG) of a planar point set is given. The expected running time of the algorithm is linear for a point set in a unit square when the points have been generated by a homogeneous planar Poisson point process. The worst-case running time is quadratic on the number of the points. The algorithm proceeds in two steps. First, a supergraph of the RNG is constructed with the aid of a cell organization of the points. Here, a point is connected by an edge to some of its nearest neighbours in eight regions around the point. The nearest region neighbours are chosen in a special way to minimize the costs. Second, extra edges are pruned from the graph by a simple scan.     

Feature Point Tracking for Incomplete Trajectories

A new algorithm is presented for feature point based motion tracking in long image sequences. Dynamic scenes with multiple, independently moving objects are considered in which feature points may temporarily disappear, enter and leave the view field. This situation is typical for surveillance and scene monitoring applications. Most of the existing approaches to feature point tracking have limited capabilities in handling incomplete trajectories, especially when the number of points and their speeds are large, and trajectory ambiguities are frequent. The proposed algorithm was designed to efficiently resolve these ambiguities. Correspondences between moving points are established in a competitive linking process that develops as the trajectories grow. Appearing and disappearing points are treated in a natural way as the points that do not link. The proposed algorithm compares favorably to efficient alternative algorithms selected and tested in a performance evaluation study. Received: June 8, 1998; revised November 18, 1998     

模糊PID控制技术研究发展回顾及其面临的若干重要问题

简要回顾了模糊PID控制器的研究发展状况,系统地归纳了模糊PID控制器的 基本形式.讨论了阻碍模糊控制技术发展的几个重要理论问题,包括系统化设计方法,控制 性能优于传统PID控制器的模糊控制器设计,模糊控制器非线性逼近能力及其量化指标,模 糊系统的规则爆炸.文中试图对有关问题的内涵进行重新定义,以及探求解决难题的关键. 如将模糊系统中的"规则爆炸"问题更正为"参数效率"问题.最后,提出了对模糊控制 理论与技术发展的思考,其中特别强调应把"简单性"作为智能控制器设计中的基本策略.     

半直觉模糊图与应用

将对象作顶点集,用直觉模糊数刻画对象间的相关性和不相关性表示成直觉模糊边;建立了半直觉模糊图模型。定义了半直觉模糊图的生成子图、度、路径、相关截图、序关系、最大生成树等概念。给出基于半直觉模糊图的聚类分析算法,分析了算法的复杂度。结合经典实例作了基于半直觉模糊图的聚类分析,结果显示基于半直觉模糊图的聚类分析算法复杂度低于一般直觉模糊聚类算法。高效实用且自动化程度高。     

平面轮廓的分段与识别技术

采用十一点法计算轮廓的近似曲率，该处比普通的三点法具有更强的抗噪性，提出利用曲率差分图进行特征点识别，并根据曲率差分图中轮廓特征点对应的波形特点，一次性将特征点识别为角点与切点，最后，对特征点间的线段类型加以识别，并进行连接调整和圆弧的分割与融合，从而较精确地实现了平面轮廓的分段及识别。     

基于改进图卷积神经网络的配网线路故障诊断

随着电网的不断扩容,系统结构越来越复杂,多故障频发,而多故障是故障诊断的关键和难点。为解决故障处理数据量大,需要快速、准确地诊断电网故障的问题,本文提出了一种基于模糊优化图卷积神经网络的配网故障诊断模型。首先处理采集的配网故障线路的特征数据;其次,搭建基于图卷积神经网络的故障诊断模型,利用模糊理论建立配电网故障的隶属函数;最后利用训练好的模型进行配网故障诊断。仿真结果表明,模糊优化图卷积神经网络对多故障诊断的准确率高于卷积神经网络以及其他方法,本文方法做出的诊断结果更加精确,综合诊断效果最好。     

Multi-scale sparse feature point correspondence by graph cuts

This paper presents a global stereo sparse matching technique based on graph cut theory to obtain accurate correspondence of stereo point pair in vision measurement applications.First, in order to obtain accurate location of feature points, a new multi-scale corner detection algorithm is proposed where wavelet coefficients are used to determine feature points by calculating an auto-correlation matrix.A sparse graph is constructed based on the feature points according to the graph cut theory.Then, the featur...