基于智能点的工厂模型设计与实现

在参考连接点和对偶点技术的基础上,提出了智能点的概念,避免了以前方法在模型统一性和结构简单性方面的缺陷,对智能点进行了详细的描述和分类,并提出了智能组的概念,基于智能点和智能组,提出了单根管线图的深度优先搜索算法。     

基于点重要性判断的点云简化

为了有效保持散乱点云的显著几何特征,提高点云简化的精度和效率,提出一种点重要性判断点云简化方法.首先,计算点云中点的重要性,并根据重要性提取特征点;然后,采用八叉树算法对非特征点进行简化,从而保留点云的主要细节特征,实现点云简化处理;最后,通过对公共点云和文物点云数据模型的简化实验来验证该点云简化方法.结果表明,该点重要性判断点云简化方法可以在有效保持点云细节几何特征的同时,实现点云的有效简化,是一种快速、高精度的点云简化方法.     

图象特征点集配准的加权相关迭代算法

图象配准是计算机视觉中目标识别的一种基本方法 .其目的是在待识别图象中寻找与模型图象的最佳匹配 .该文以传统的 U meyam a点集相关度量为基础 ,结合 Procrustes正规化方法 ,通过引入加权矩阵 ,以得到新的相关度量函数 ,进而提出了一种图象特征点集匹配的新方法 ,解决了传统方法要求点集维数相同的缺点 .经过迭代运算 ,对存在几何失真 ,且维数不同的两点集可得到精确配准 .文中给出的点集配准结果说明 ,当两点集维数相同时 ,该方法不仅与传统的点集相关法一样 ,均可达到精确配准 ,而且该方法对维数不同的点集也可实现精确配准     

四路摄像头协同多重触控系统标定方法

四路摄像头协同多重触控系统通过拟合四路摄像头得到触点方向线,以解决多触点遮档问题。针对该原理中触点方向线构建问题,提出查找表法、消失点法和立体标定法3种系统标定方法,以实现多触点精确定位。查找表法在交互区边框上设定标定点,通过插值查找对应标定点,以光心和标定点构建触点方向线;消失点法利用触点方向上的消失点和光心构建触点方向线;立体标定法利用极线几何反投影射线构建触点方向线。误差分析实验结果表明,查找表法适用于较小尺寸平台,消失点法适用于大尺寸平台,立体标定方法精度有待提高。     

基于特殊角点和基点定位在线掌纹的方法

本文简述了目前在线掌纹定位的主要方法,提出一种通过寻找特殊角点和基点实现在线掌纹的定位。该方法基于掌纹的形状特点,先定位特殊角点,再通过统计块点方向和求梯度确定感兴趣的纹线上的一点,最后用灰度值跟踪方法确定特殊基点。该定位方法减少了手掌张开强度因素造成的影响,并取得了理想的实验结果。     

基于RBF的点模型布尔运算

由已有模型构造新的模型是数字娱乐领域的重要造型方法,提出了一种基于RBF（径向基函数）的点模型布尔运算方法。该方法首先将输入点模型进行空间八叉树划分,然后采用RBF将点模型隐式化生成隐函数表示的符号距离场,最后通过隐式曲面的布尔运算方法剔除多余的点,得到需要的新点模型。对点模型进行空间八叉树划分加速了点模型的RBF隐式化过程。该方法可由多个简单的点模型快速构造出复杂的点模型。     

Computation of a simple bifurcation point using one singular value decomposition nearby

Using an extended system to locate a simple bifurcation point via an iterative method usually requires a good choice of an initial point as well as several auxiliary vectors. The method we propose here requires a good choice of an initial point only. Our method is based upon an analysis of singular vectors of a singular value decomposition of a Jacobian matrix near the simple bifurcation point, which leads to the automatic determination of a certain type of auxiliary vectors in terms of the initial point. Numerical implementation via a Newton-like method is discussed and examples are provided.     

基于时空约束密度聚类的职住地识别方法

为了从移动终端位置数据中精准识别居民职住地,提出了一种基于时空约束密度聚类的职住地识别方法。首先,利用基于K-means的DBSCAN（density-based spatial clustering of applications with noise）时空驻点聚类过程将居民多天的原始轨迹点分成不同的时空驻点簇。然后,利用基于速度阈值的停留点簇和移动点簇识别过程将居民的每一个时空驻点簇区分为停留点簇或移动点簇。接着,利用基于K近距离的DBSCAN重要停留点聚类过程将居民的停留点分成不同的重要停留点簇。最后,利用基于KD-tree优化的KNN（K-nearest neighbor）职住地识别过程将居民的每个重要停留点识别为工作地、居住地、职住同一区域或兴趣地点区域。实验结果表明,该方法的每个过程都是合理有效的,并且最终的职住地识别效果要优于时间阈值法、累加时间法和信息熵法。     

无人机摄影测量点云道路自适应提取

在无人机摄影测量中, 针对传统的地面点云提取方法对图像点云数据中的道路提取适应性较差的问题, 本文提出了一种无人机摄影测量点云道路自适应提取方法. 首先, 根据点云的空间几何特征将点云划分为3个类别; 然后, 针对非道路的点云类别采取相应的方法进行剔除; 最后, 对经过自适应提取方法得到的点云数据进行滤波平滑和基于颜色的区域生长分割处理. 实验结果表明, 该方法提取的道路点云的I类误差为4.97%, II类误差为1.14%. 该方法能够有效地提取目标道路路面, 提高了无人机摄影测量工程应用中点云数据处理的效率.     

基于深度学习特征点法的单目视觉里程计

针对特征点法的视觉里程计VO中光度、视点变化对特征点提取稳定性降低的不利影响,提出一种基于深度学习特征点法的单目VO方法。采用自监督深度学习网络训练得到DSP特征点检测器。首先使用亮度非线性逐点调整方法对训练图像进行光度调整;然后使用非极大值抑制方法剔除冗余DSP特征点,改进最邻近方法得到双向最邻近方法,解决特征点匹配问题;最后建立最小化重投影误差方程求解优化位姿及空间点参数。在Hpatches、Visual Odometry数据集上进行验证,实验结果表明:DSP特征点检测器增强了特征匹配对光度、视点变化的鲁棒性;无后端优化的条件下,本方法定位均方根误差比ORB方法明显降低,且保证了系统实时性,为特征点法的VO提供新的解决思路。     

基于边缘关联点云的激光雷达与相机外参标定方法

激光雷达的点云和相机的图像经常被融合应用在多个领域。准确的外参标定是融合两者信息的前提。点云特征提取是外参标定的关键步骤。但是点云的低分辨率和低质量会影响标定结果的精度。针对这些问题,提出一种基于边缘关联点云的激光雷达与相机外参标定方法。首先,利用双回波提取标定板边缘关联点云;然后,通过优化方法从边缘关联点云中提取出与实际标定板尺寸大小兼容的标定板角点;最后,将点云中角点和图像中角点匹配。用多点透视方法求解激光雷达与相机之间的外参。实验结果表明,该方法的重投影误差为1.602px,低于同类对比方法,验证了该方法的有效性与准确性。     

基于固定电压法结合P＆O法MPPT控制策略研究

对光伏发电控制系统提出了一种新型的最大功率点跟踪（MPPT）控制方法。即在外界环境或负载突变时,先采用固定电压法将光伏阵列的工作点调整到最大功率点附近,以保证跟踪的快速性;并且引入新型的扰动观察法,对最大功率点的稳态特性进行优化,可有效减小光伏阵列的输出功率在最大功率点的振荡现象。通过仿真验证了上述方法的有效性。     

Automatic least-squares projection of points onto point clouds with applications in reverse engineering

A novel method for projecting points onto a point cloud, possibly with noise, is presented based on the point directed projection (DP) algorithm proposed by Azariadis P., Sapidis N. [Drawing curves onto a cloud of points for point-based modelling. Computer-Aided Design 2005; 37(1): 109–22]. The new method operates directly on the point cloud without any explicit or implicit surface reconstruction procedure. The presented method uses a simple, robust, and efficient algorithm: least-squares projection (LSP), which projects points onto the point cloud in a least-squares sense without any specification of the projection vector. The main contribution of this novel method is the automatic computation of the projection vector. Furthermore, we demonstrate the effectiveness of this approach through a number of application examples including thinning a point cloud, point normal estimation, projecting curves onto a point cloud and others.     

点云数据在深度学习中表示方法的研究

为了实现在深度学习中能够端到端表示点云模型,提出基于八叉树和K-D树（OctKD）的点云数据表示方法。该方法将无组织的点云转换为体素空间,在体素空间对三维模型进行八叉树剖分,改进了八叉树编码方式;构建节点间的邻接关系,在GPU端并行构建八叉树;为了克服八叉树编码检索效率低的问题,采用三维K-D树索引单个三维空间点。实验结果表明该方法能够真实反映模型本身的细节特征,提高了点云模型的构造时间和检索效率。这种新的数据结构实现将点云转换为卷积神经网络可以接收的数据形式。     

Fast Hybrid Approach for Texturing Point Models

We present three methods for texturing point models from sample textures. The first method, the point parameterization method, uses a fast distortion‐bounded parameterization algorithm to flatten the point model's surface into one or more 2D patches. The sample texture is mapped onto these patches and alpha blending is used to minimize the discontinuity in the gaps between the patches. The second method is based on neighborhood matching where a color is assigned to each point by searching the best match within an irregular neighborhood. The hybrid method combines the former two methods, capitalizing on the advantages of both. The point parameterization method is used first to color most of the points, and the point neighborhood‐matching method is then applied to the points belonging to the gaps between the parameterized patches to minimize the discontinuity. We opt for fast texture generation, while some discontinuities may appear in the gaps of anisotropic textures.     

Registration of point clouds using sample-sphere and adaptive distance restriction

Registration of point clouds is a fundamental problem in shape acquisition and shape modeling. In this paper, a?novel technique, the sample-sphere method, is proposed to register a pair of point clouds in arbitrary initial positions. This method roughly aligns point clouds by matching pairs of triplets of points, which are approximately congruent under rigid transformation. For a given triplet of points, this method can find all its approximately congruent triplets in O(knlog?n) time, where n is the number of points in the point cloud, and k is a constant depending only on a given tolerance to the rotation error. By employing the techniques of wide bases and largest common point set (LCP), our method is resilient to noise and outliers. Another contribution of this paper is proposing an adaptive distance restriction to improve ICP (iterative closest point) algorithm, which is a classical method to refine rough alignments. With this restriction, the improved ICP is able to reject unreasonable corresponding point pairs during each iteration, so it can precisely align the point clouds which have large non-overlapping regions.     

复杂背景下红外图像弱小点目标自动识别仿真

传统识别方法受到低信噪比、低对比度、缺乏弱小点目标的形状及纹理信息等因素影响,尤其在复杂背景下,弱小点目标自动识别准确率较低,针对此问题,提出一种基于BEMD(二维经验模态分解算法)的红外图像弱小点目标自动识别方法,根据待识别图像的频谱特性,并结合分频段处理方式。对比了不同滤波器的性能,并建立了图像滤波器组,采用滤波器组将弱小点目标图像分解到不同子频域中;对子频段图像进行罗宾逊滤波处理,提取弱小点目标。采用多层经验模态分解算法对原始弱小点目标图像输入函数分解为二维本征模态函数,通过微分计算来获取原始图像与背景区域之间的差,分割出弱小点目标区域。通过局部逆熵分割弱小点目标区域的高频信息来获取各个模态函数的弱小点目标识别结果。实验结果表明,所提方法能够高效且准确地提取出弱小点目标,更好地抑制复杂背景。     

A comparative analysis of two point cloud volume calculation methods

A spatial body object morphology analysis basically consists of the volume parameter. However, the challenge lies in the volume calculations of irregular objects. A point cloud slicing method (PCS_l) is proposed in this study. In this method, a uniformly spaced sequential slicing process is first conducted in a specific direction through the point cloud of a spatial object obtained via three-dimensional laser scanning (3DLS). A series of discrete point cloud slices corresponding to the point cloud body are then obtained. Subsequently, the outline boundary polygons of the point cloud slices are searched one by one in accordance with the slicing sequence and the areas of the polygons. The point cloud slice is also calculated. Finally, the individual point cloud section volume is calculated throughout the slicing areas and the adjacent slicing gap. In this way, the total volume of the scanned spatial object can be calculated by summing up the individual volumes. According to the results and analyses of the calculated examples, the slice-based volume-calculating method for the point cloud of irregular objects obtained via 3DLS is correct and has a concise process, reliable results, efficient calculation methods and a controllable accuracy. The PCS_l algorithm proposed in this study effectively works to overcome the disadvantages of the point cloud segmenting method. It also provides a quick and accurate solution to the volume calculation of a point cloud of a spatial object obtained via 3DLS. This method provides a good solution for volume calculations of irregular objects.     

飞行时间三维相机的多视角散乱点云优化配准方法研究

针对目前基于飞行时间(TOF)原理的三维相机实现物体完整表面的三维点云重建过程中,多视角散乱点云配准精度低的问题,本文提出了一种多视角散乱点云优化配准方法。该方法通过构建一个目标功能函数,并结合相邻点云的变换矩阵对该目标函数进行最小化求解,直接获取任意位置的点云到基准坐标系的绝对变换矩阵,避免了对相邻点云的变换矩阵进行累积而引起误差的累加。实验结果表明,该方法提高了多视角点云配准的精度,同时增强了物体点云模型重建的效果,在三维曲面重建中具有较强的实用性。     

信号奇异点的分步检测法

针对两种常见的信号奇异点：脉冲型奇异点和阶跃型奇异点,证明信号的奇异点与信号小波变换的最值有关,如果适当选择小波基函数,那么信号的脉冲奇异点将对应于小波基函数的最值点,而信号的阶跃奇异点将对应于小波基函数的原函数的最值点。据此,设计了一个新的基于小波变换的信号奇异点分步检测法（Hierarchical Singular Point Detection based on Wavelet Transform,HSPDWT）,该方法的特点是根据脉冲奇异点和阶跃奇异点的不同特征分两步从信号中提取奇异点。仿真及真实信号上的实验证明了HSPDWT的可行性和有效性。