一种基于文法演化自动拟合非线性数据的蜂群算法

将文法演化方法引入蜂群算法, 基于上下文文法定义多种数学函数, 提出了自动拟合非线性数据的蜂群算法BCGE, 并给出了通过基因截断、基因增补及利用领域知识加速BCGE的方法。基于文法演化的BCGE比基于文法规划的其他算法更为简洁, 且存储基因型所需的空间也远比其他算法存储语法树所需的空间少。通过五个测试函数的实验表明, BCGE能够有效地拟合非常复杂的非线性数据。     

遗传规划和进化策略混合算法及应用

文中把函数拟合建模看作是模型结构和参数的优化搜索过程,将遗传规划和进化策略结合起来对函数拟合的结构和参数共存且相互影响的复杂解空间进行全局最优搜索实现拟合结构和参数的共同识别。克服了传统的函数拟合完全依赖于数据、精度低、结构与参数分别确定这样一“串行”计算结构等缺陷。实验数据表明,该方法得到的拟合函数比传统方法得到的拟合函数,具有较高的精度和推广预测能力。     

Interactive pattern recognition of blood cells in malignant lymphomas

A data set consisting of blood cells from patients with Hodgkin's disease or non-Hodgkin lymphoma was explored using the interactive system for statistical pattern recognition ISPAHAN. The results show the usefulness of statistical pattern recognition in hematology.     

计算机曲线拟合求解析氢反应的动力学参数

应用电化学稳态极化方法,研究1M NaOH电解液中镍硫合金电极表面析氢反应电催化活性与沉积层中含硫量之间的关系.根据析氢反应机理,用理论电化学极化方程拟合实验稳态极化曲线,得到不同含硫量的镍硫合金电极表面氢原子吸附和脱附反应的速率常数、反应对称系数等动力学参数.     

基于内容的交互式感性图象检索

随着信息化社会的到来及信息高速公路计划的实施,人们越来越多地接触到大量的图象信息,因此基于内容的图象检索已经成为当前的一个热门研究课题,并在多媒体数据库、电子图书馆、商标管理、医疗图象管理、公安系统、卫星图象管理等方面得到广泛应用。然而,大多数基于内容的图象检索系统主要是通过图象多维物理特征的相似性匹配来进行查询,而对于用户的爱好、情感等主观或感性化的因素则考虑较少。为了弥补这方面的不足,提出了一种基于内容的交互式感性图象检索方法。该方法采用交互式进化算法,并通过人机交互的方式,来将用户的直觉、情感等感性化的因素融入到进化过程,以便进行图象的交互式在线检索;针对在检索过程中,因进化的时间可能较长和因需要用户确定的适应度值较多而产生的用户疲劳问题,采用神经网络离线学习的方法来减轻用户疲劳,从而实现了根据用户的情感和基于图象内容的图象检索,并取得了较好的实验结果。     

Interactive videodiscs for science education

Abstract Videodisc technology allows the marriage of a uniquely versatile visual medium with interactive computer-assisted learning packages and the combination can provide a very powerful educational resource. This review paper explores the teaching and learning potential of interactive video (IV), with reference to science courses, and describes some of the programmes currently in use.     

基于局部-非局部交互卷积的3D点云分类

现阶段点云分类研究已被广泛应用于机器人操作、自主驾驶和虚拟现实等多个领域,提取既丰富又具有高判别能力的特征是3D点云分类的关键.为此,文中设计基于局部-非局部交互卷积的3D点云分类算法,改善点云的特征提取.首先,构造局部-非局部交互卷积模块,在获取局部相似特征和非局部相似特征的基础上,采用交互增强,缓解单个邻域在表示封闭区域时存在的冗余问题,增强网络的层次性和稳定性,同时也缓解网络的退化问题.然后,以该模块为基本单元构建卷积神经网络.最后,采用自适应特征融合,充分利用不同层次的特征,实现3D点云的分类.在ModelNet40、ScanObjectNN基准数据集上的实验表明,文中算法性能较优.     

基于安德森加速的快速 B 样条拟合算法

曲线拟合技术已被广泛地应用于图像处理、工程实验等领域。其中,B 样条曲线拟 合是曲线拟合中最常见的方法,它具有局部性好、连续性好等优点,但拟合精度一般较低。在实 际应用中,B 样条曲线拟合对于精度和速度的要求都较高。为了提升平面 B 样条曲线拟合速度, 将安德森加速的想法应用到曲线拟合的方法之中,提出一种基于安德森加速的拟牛顿方法。首先 设定一个初始形状,然后根据初始形状找到其每个数据点的投影点的位置参数,然后利用安德森 加速计算出控制点的相应位置,迭代进行以上 2 步,直到结果收敛。实验结果表明,该方法在收 敛速度和迭代时间上均优于其他方法。     

基于包络拟合法的FPGA超声测距系统设计

针对普通超声测距系统精度低、速度慢的问题,提出了一种全硬件实现的FPGA超声测距系统。将最小二乘法的二次曲线拟合算法应用于超声回波包络拟合,完成回波信号的数字信号处理和距离的测量。采用硬件描述语言在Altera公司的EP2C70F896C6上实现,在4 m范围内测距误差小于±1 mm。该系统具有精度高、运算速度快的特点,并具备很强的功能扩展性,可扩展到超声探伤、超声成像等领域。     

GPS高程拟合在水利测量中的应用研究

常见高程系统包括大地高、正高及正常高系统,由于 GPS 测量的大地高在实际测量中的应用十分有限,须采用高程拟合技术,将大地高转换成为正常高,但转换过程中会出现高程异常,因此采用拟合方程对已知高程进行异常处理。高程拟合方法包括多项式与多面函数拟合法,在 GPS 拟合原理基础上,对二次多项式、 三次多项式、多面函数 3 种拟合方法的高程异常及拟合差值进行分析计算,并将拟合方法应用于地形起伏差异不同的 2 个测区,对 3 种拟合方法的精度评定进行比较分析。结果表明：三次多项式拟合方法的精度要优于二次多项式和多面函数的拟合精度,说明三次多项式拟合更适用于水利测绘工程,且通过三次多项式拟合方法可减少水利测量工作,提高测量精度和工作效率。     

基于改进聚类算法的传感器非线性数据拟合研究

针对传感器输出电信号与物理量之间存在的非线性问题,提出了一种新的曲线拟合方法。该方法通过将改进的粒子群优化算法,混沌搜索法和改进的模糊C-均值算法相结合,对实验数据搜索聚类中心点,然后利用分段线性逼近对传感器输入输出关系进行拟合。介绍了粒子群优化算法和模糊C-均值算法,给出了相关的公式推导过程,对传感器输出电信号与物理量的对应关系进行曲线拟合,最后将该方法应用于电涡流智能传感器。实验结果表明,该方法精度高、可靠性好,具有较强的自适应性和快速性,能够更准确的将电信号转换为物理量。     

基于二次曲面拟合的二维传感器数据处理技术研究

针对温度、湿度传感器在测量系统中相互干扰所导致的测量误差问题,提出了基于二次曲面拟合理论的数据处理模型,实验以温湿度集成传感器为例,采用传感器标定系统对二维传感器进行多点温度、湿度标定,基于二次曲面拟合方程对数据进行处理,结果表明:该方法有效地提高了温度、湿度传感器的测量精度,具有很高的实用价值。     

一次性注射器针头反表识别系统

介绍一种能正确识别一次性注射器针头在装配过程中是否反表的系统。系统由面阵光源、普通数字摄像机、运动控制卡、步进电机、针排夹具和接近开关等组成;采用虚拟仪器的IMAQ Vision模块进行编程,对摄像机采集到的针头部分图像进行自动阈值处理、二值化、目标定位、模板匹配等图像分析处理;由于针排与针可能存在一定的倾斜,所以,在模板匹配前需要对采集到的图像进行校正;通过计算模板与搜索图之间的相似度对针头反表进行识别,并进行显示报警。     

高斯拟合亚像素边缘检测算法

针对传统边缘检测算法的定位精度低、对噪声敏感等缺点,提出基于函数曲线拟合的亚像素边缘检测算法——梯度方向高斯曲线拟合亚像素定位算法。该方法首先在边缘附近选取一系列点,求得这些点的灰度值,进而求得这些点的梯度值,然后运用高斯曲线来对这些点的梯度值进行拟合,最后通过拟合曲线求得高斯曲线的对称轴位置即为亚像素位置。实验表明该算法能够很好地实现亚像素定位,通过与其他两种亚像素定位算法的比较,得出该算法运行时间较短,效率较高。     

Intelligent sensor system for discrimination of material type using neural networks

Neural network based classification of material type even with the variation in the sensor parameter is investigated in this paper. The sensor is developed by means of a lightweight plunger probe and an optical mouse sensor. An experimental prototype was developed which involves bouncing or hopping of the plunger based impact probe freely on the plain surface of an object under test. The experiment is conducted to obtain the bouncing signals for plain surface of an objects kept at different distances from the probe. During the bouncing of the probe, time varying signals are generated from optical mouse that are recorded in data files on PC. Some dominant unique features are then extracted using signal processing tools to optimize neural network based classifier. The time and features of bouncing signal are related to the material type, and each material has a unique set of such properties. It is found that the sensor system is intelligent due to its ability to classify the material type even with the variation in the sensor parameter (distance between the sensor probe and plain objects). The classifiers are developed using two neural networks configurations, namely a well-known Multi-layer Perceptron Neural Networks (MLP NN), and Radial Basis Function Neural Networks (RBF NN). MLP NN and RBF NN models are designed to maximize accuracy under the constraints of minimum network dimension.The optimal parameters of MLP NN and RBF NN models based on various performance measures that include percentage classification accuracy (PCLA) on the testing data, and area under Receiver Operating Characteristics (ROC), and are determined. For the sensor data set, the PCLA of both the classifiers are found reasonable consistently in respect of rigorous testing using different data partitions. The areas under the ROC curves are close to unity. Performances of the two classifiers have been compared. It has been found that the RBF NN is more robust to noise, and epochs required for training are very less as compared to that for MLP NN.     

面向交互式动画创作结构化剧本的生成

为了减少传统动画制作中手工劳动,实现高效的动画剧本创作,提出了一种面向交互式动画创作的结构化动画剧本生成模型并加以实现。通过一个实际样例阐述了该模型在剧本生成过程的具体应用,并对其结果进行分析,验证了模型的有效性。     

FILiC:一种CUDA上的交互型库函数框架

CUDA是NVIDIA公司推出的GPU编程模型,它为高效利用GPU计算能力提供了强大的支持。但CUDA线程无法直接访问I/O设备、网卡等外围设备,在CUDA线程与外围设备的交互功能方面,目前CUDA的支持十分有限,仅支持非实时的屏幕打印(printf)。因此提出了一种交互型库函数框架FILiC,它通过设备和主机之间的巧妙交互,高效实现了CUDA线程实时的较完整I/O等函数;并且该框架具有很好的可扩展性,CUDA程序员或者编译器开发者可基于该框架按需求开发新的CUDA线程交互功能。     

Knot calculation for spline fitting via sparse optimization

Curve fitting with splines is a fundamental problem in computer-aided design and engineering. However, how to choose the number of knots and how to place the knots in spline fitting remain a difficult issue. This paper presents a framework for computing knots (including the number and positions) in curve fitting based on a sparse optimization model. The framework consists of two steps: first, from a dense initial knot vector, a set of active knots is selected at which certain order derivative of the spline is discontinuous by solving a sparse optimization problem; second, we further remove redundant knots and adjust the positions of active knots to obtain the final knot vector. Our experiments show that the approximation spline curve obtained by our approach has less number of knots compared to existing methods. Particularly, when the data points are sampled dense enough from a spline, our algorithm can recover the ground truth knot vector and reproduce the spline.     

综述MATLAB与VC++的交互编程

在通信等工程领域的计算机仿真中,VC和Matlah的交互编程可以结合两者的优点,提高仿真效率。文中介绍了Matlah与VC 之间交互编程的几种方法,包括调用Matlab引擎、C语言mex程序、mat文件操作程序等方法,并特别介绍了Matlab Visual Studio Add-in的使用。同时分析了各种方法的优缺点和适用场合,提出了适合于算法仿真的交互编程方法。     

Array algebra polynomial fitting for image segmentation

Segmentation of digital images is a key problem for automation of object recognition and object reconstruction. Area based segmentations, like region growing and split-and-merge, may use similarity with respect to some grey level property as predicate for appending pixels to a region. Regions of particular interest, for example roofs of buildings in aerial photographs, are seldom homogeneous with respect to grey level. There are often position dependent trends in the grey values of a region, that is; the image is piecewise smooth. These trends may be modelled by functional fitting. This turns image segmentation into a computer intensive least squares problem. In this article, a split-and-merge procedure using array algebra is proposed to make polynomial fitting faster. The splitting is two to three times faster than using regular matrix algebra. For 3D images it is four to nine times faster. The merging is performed without accessing the image.This study was financed by the Swedish Council for Building Research (BFR).