用于线段特征提取的改进Hough变换

线段是符号的一个具有旋转、平移和尺度不变的稳定性特征,正确提取符号的线段特征对于提高符号识别系统的识别率有很重要的意义。针对已有的基于Hough变换的线段提取算法的缺点,该文提出了一种用于线段特征提取的改进Hough变换算法。通过采用“多对一”映射;将Hough变换的投票过程和线段参数的检测过程融为一体;动态管理算法所需的临时存储空间等手段,使该算法具有较好的计算复杂度和空间复杂度。针对数字图像的量化特点,精心设计了用于检测在直线上点的条形区域,从而大大地降低了噪声对线段参数检测的影响,使该算法具有较好的检测性能和鲁棒性。实验表明,该文算法能正确提取出线段的端点坐标及其长度。     

命题μ-演算局部模型检测高效算法设计

命题μ-演算局部模型检测算法中,目前最好的算法的时间复杂度与不动点算子交替嵌套深度d呈指数关系。针对命题μ-演算局部模型检测算法的计算过程进行分析,得到迭代计算的中间迭代值间满足的一组偏序关系,然后利用该偏序关系设计了一个局部模型检测算法,算法时间复杂度的指数部分为d/2,大大提高了算法的计算效率。     

基于改进的多分辨率Hough变换的直线检测方法

该文提出了一种基于改进的多分辨率Hough变换的直线检测方法。对多分辨率图像由粗到细地循环处理，首先利用全局累加器序列对最小的图像做Hough变换得直线参数，利用该直线参数估计下一循环图像中的直线参数；建立了一种简单对应关系，将得到的直线参数的估计范围转换成图像空间范围；然后在图像空间范围内利用局部累加器对较大图像做Hough变换检测直线。最后还给出了线段端点检测、长度计算及连续性考察的方法。实验表明，该方法可以加快直线检测速度、节省存储空间，端点检测方法具有一定的鲁棒性。     

噪声环境下基于局部不确定性度量的多直线检测算法

针对噪声环境下复杂工件多直线测量中存在的检测精度下降问题,提出了一种基于局部不确定性度量的直线检测算法。该算法首先将目标区域分块,在每个子块内,建立点属于某直线不确定性度量概率模型,按照两点组合原理和Bayesian法则,计算任两点所确定的直线的累积不确定度量,通过对参数空间软投票检测直线。算法具有较强的抗噪能力和快速性。实验结果表明,该算法在较高噪声（方差为0.06）时,检测精度误差小于1‰,检测正确率仍可以达到90%以上,且其检测时间是单纯不确定度量直线检测方法的1/2, 比传统Hough变换算法快4~5倍。     

直线拟合快速实现的一种新算法

对三种基于残差准则的直线拟合方法进行分析并提出了一种直线拟合快速实现的新算法——参数逼近法；算法以变步长对斜率和截距进行逼近，降低了利用计算机进行直线拟合的复杂度；算法不仅容易实现，收敛速度快，而且可以根据需要自由控制拟合精度；另外，算法对3种准则的拟合均适用，从而将3种直线拟合法统一于一个体系之下。     

基于相位编组图像分块的快速Hough变换直线检测

在分析Hough变换直线检测算法和相位编组法直线检测算法的基础上，针对这两个直线检测算法的不足，结合它们的优点，设计并实现了基于相位编组图像分块的快速Hough变换直线检测算法，对算法进行了详细描述和算法优点分析，并通过实验验证了算法的有效性，实验表明所设计的直线检测算法运算速度快，参数易于选择，鲁棒性强，有一定的应用价值。     

基于地统计学的空间离群点检测算法的研究

针对传统空间离群点检测算法构建邻域时参数选择困难,处理高维数据的时间复杂度较高等问题,提出了一种基于地统计学的空间离群点检测算法。该算法将空间自相关理论引入空间离群检测中,首先利用3σ规则识别全局离群点,然后利用Delaunay三角网构建空间邻域,将邻域节点均值代替全局离群点,最后使用局部Moran’ I作为空间异常的度量方法。仿真结果表明,该方法不需要选择参数,鲁棒性较强,检测率较高、误警率较低。     

基于正则表达式的深度包检测算法

在深入分析了DFA状态数对算法性能影响的基础上，提出了一种新的基于正则表达式的深度包检测算法，该算法保证在任意有限的系统资源下算法的时间复杂度空间复杂度最小。在Linux下实现了该算法，并对基于L7-filter模式集合的网络数据包进行了大量检测实验。结果表明，与已有的正则表达式算法比较，该算法的时间复杂度和空降复杂度最小。     

基于自适应LBP的视频文本检测算法

针对LBP算法自适应性弱及复杂度高的问题,提出一种基于自适应LBP算子的视频文本检测算法。该算法利用全局及局部的像素灰度均差决定自适应阈值大小,能最大限度去除复杂背景,自适应性较强。给出基于近似圆的掩模算法来覆盖多方向种子生长区,降低其复杂度。实验结果表明,该算法在复杂背景下能取得较高的检测率,具有良好的自适应性和实用性,复杂度较低。     

基于机器视觉的织物检测方法的研究

对布匹表面质量的检测方法进行了研究,并提出了一种新的检测思路.与其它基于纹理的布匹缺陷处理算法不同,该方法的关键步骤就是采用了一种自定义的局部二值化操作,并经实验验证取得了较为理想的处理效果.它具有较为理想的缺陷参数提取效果,较低的误检率和较低的算法复杂度.     

A hybrid of electromagnetism-like mechanism and back-propagation algorithms for recurrent neural fuzzy systems design

This article introduces a novel hybrid evolutionary algorithm for recurrent fuzzy neural systems design in applications of nonlinear systems. The hybrid learning algorithm, IEMBP-improved electromagnetism-like (EM) with back-propagation (BP) technique, combines the advantages of EM and BP algorithms which provides high-speed convergence, higher accuracy and less computational complexity (computation time in seconds). In addition, the IEMBP needs only a small population to outperform the standard EM that uses a larger population. For a recurrent neural fuzzy system, IEMBP simulates the ‘attraction’ and ‘repulsion’ of charged particles by considering each neural system parameters as a charged particle. The EM algorithm is modified in such a way that the competition selection is adopted and the random neighbourhood local search is replaced by BP without evaluations. Thus, the IEMBP algorithm combines the advantages of multi-point search, global optimisation and faster convergence. Finally, several illustration examples for nonlinear systems are shown to demonstrate the performance and effectiveness of IEMBP.     

Quantitative temporal logics over the reals: PSpace and below

In many cases, the addition of metric operators to qualitative temporal logics (TLs) increases the complexity of satisfiability by at least one exponential: while common qualitative TLs are complete for NP or PSpace, their metric extensions are often ExpSpace-complete or even undecidable. In this paper, we exhibit several metric extensions of qualitative TLs of the real line that are at most PSpace-complete, and analyze the transition from NP to PSpace for such logics. Our first result is that the logic obtained by extending since-until logic of the real line with the operators ‘sometime within n time units in the past/future’ is still PSpace-complete. In contrast to existing results, we also capture the case where n is coded in binary and the finite variability assumption is not made. To establish containment in PSpace, we use a novel reduction technique that can also be used to prove tight upper complexity bounds for many other metric TLs in which the numerical parameters to metric operators are coded in binary. We then consider metric TLs of the reals that do not offer any qualitative temporal operators. In such languages, the complexity turns out to depend on whether binary or unary coding of parameters is assumed: satisfiability is still PSpace-complete under binary coding, but only NP-complete under unary coding.     

An automatic time-stepping algorithm for dynamic problems

The paper discusses algorithms for automatic computation of time steps in discrete integration of dynamic problems. Objective criteria for performance and guidelines for making an algorithm practically applicable are suggested. A set of parameters for characterizing the dynamic response of a system is proposed, this includes a ‘current frequency’, a ‘current period’, and a ‘dynamic stiffness parameter’. An algorithm for automatic computation of the time step based on the current period is outlined. Aspects of adapting this algorithm to special types of applications are discussed, this involves use of a ‘tuning function’. Several examples of applications are given.     

Breaking paragraphs into lines

This paper discusses a new approach to the problem of dividing the text of a paragraph into lines of approximately equal length. Instead of simply making decisions one line at a time, the method considers the paragraph as a whole, so that the final appearance of a given line might be influenced by the text on succeeding lines. A system based on three simple primitive concepts called ‘boxes’, ‘glue’, and ‘penalties’ provides the ability to deal satisfactorily with a wide variety of typesetting problems in a unified framework, using a single algorithm that determines optimum breakpoints. The algorithm avoids backtracking by a judicious use of the techniques of dynamic programming. Extensive computational experience confirms that the approach is both efficient and effective in producing high-quality output. The paper concludes with a brief history of line-breaking methods, and an appendix presents a simplified algorithm that requires comparatively few resources.     

Efficient random walks in the presence of complex two-dimensional geometries

This paper details an efficient algorithm for particles undergoing random walks in the presence of complex, two-dimensional, solid boundaries. The scheme is developed for the simulation of vortex diffusion using the random vortex method. Both vortex blobs and sheets are handled. The relevant modifications for using the algorithm with the vorticity redistribution technique are also discussed. The algorithm is designed to be used in the framework of an existing fast multipole implementation. A measure for the geometric complexity of a body is introduced and the algorithm’s efficiency is studied as various parameters are changed for bodies of varying complexity.     

Determining the mode

The complexity of computing modes and of sorting multisets is considered. Previous lower bounds are improved and an algorithm is given to determine the mode of a multiset in a number of comparisons differing from the lower bound by only a ‘lower order term’.     

基于小波变换的半软阈值参数算法研究

小波域阈值滤波因其实现最简单,计算量最小而得到广泛的应用,但对不同信号而言,其阈值函数的选择将直接影响到滤波效果。由GaoHongYe提出的半软阈值法,因其参数实现算法的复杂度大而没有得到有效的应用。结合小波理论与模糊理论对半软阈值函数的参数提出一种新的计算算法,大大减少了算法实现的复杂度,并对其进行试验仿真,取得了良好的效果。     

Robust explicit model predictive control for linear systems via interpolation techniques

This paper presents a novel interpolation‐based model predictive control (IMPC) for constrained linear systems with bounded disturbances. The idea of so‐called ‘pre‐stabilizing’ MPC is extended by making interpolation among several ‘pre‐stabilizing’ MPC controllers, through which the domain of attraction can be magnificently enlarged. Compared with the standard ‘pre‐stabilizing’ MPC, the proposed approach has the advantage of combining the merits of having a large domain of attraction and a good behavior. Furthermore, such an IMPC problem can be solved off‐line by multi‐parametric programming. The optimal solution is given in an explicitly piecewise affine form. A simple algorithm for the implementation of the explicit MPC control laws is also proposed. Copyright © 2009 John Wiley & Sons, Ltd.     

Truncated adaptation design for decentralised neural dynamic surface control of interconnected nonlinear systems under input saturation

This paper studies decentralised neural adaptive control of a class of interconnected nonlinear systems, each subsystem is in the presence of input saturation and external disturbance and has independent system order. Using a novel truncated adaptation design, dynamic surface control technique and minimal-learning-parameters algorithm, the proposed method circumvents the problems of ‘explosion of complexity’ and ‘dimension curse’ that exist in the traditional backstepping design. Comparing to the methodology that neural weights are online updated in the controllers, only one scalar needs to be updated in the controllers of each subsystem when dealing with unknown systematic dynamics. Radial basis function neural networks (NNs) are used in the online approximation of unknown systematic dynamics. It is proved using Lyapunov stability theory that all the signals in the closed-loop system are semi-globally uniformly ultimately bounded. The tracking errors of each subsystems, the amplitude of NN approximation residuals and external disturbances can be attenuated to arbitrarily small by tuning proper design parameters. Simulation results are given to demonstrate the effectiveness of the proposed method.