非刚性部分模型与完整模型的对应关系计算

针对不同姿态下的三维等距部分模型与完整模型对应关系计算问题,提出了一种结合局部函数映射和局部流形谐波(localized manifold harmonics,LMH)算子计算三维模型特征描述符并构建模型间对应关系的新方法.首先,通过改进的Laplace算子的谱分解构造局部基产生LMH算子,并计算模型的特征描述符;其次...     

顾及异向性的局部径向基函数三维空间插值

针对三维局部径向基函数(radialbasisfunction,RBF)空间插值过程中合理设置待插值点的邻近搜索范围问题,提出了一种依据变异函数分析探索空间异向性的局部径向基函数插值方法。首先,构建采样点的协方差矩阵求解数据的三个轴向;然后,通过旋转变换将数据变换到新的坐标系下,根据地统计中的变异函数计算三个方向的变程,并将三个变程设置为局部搜索椭球体的三个轴;最后,对每个采样点构建节点RBF,通过对待插值点影响范围内的节点RBF进行线性加权组合,求出待插点的属性值。实验结果表明,该方法顾及了空间数据的异向性,计算精度高,插值结果可靠,是一种可行的顾及异向性的三维空间插值方法。     

基于最优邻域局部熵的点云精简算法

针对传统的点云精简算法中不能良好保留细节特征的问题,提出一种基于最优邻域局部熵的点云精简算法.首先利用点云局部邻域协方差矩阵的3个特征值构造的维度特征,构建局部邻域信息熵函数,其次依据局部熵值最小原则确定最优邻域,然后根据最优邻域下计算的特征值间的关系,以及局部信息熵来剔除平坦区域数据点.通过模拟数据和实例扫描数据精简...     

更充分利用独立弯曲结构的线状要素Morphing变换方法

本文提出了充分利用独立弯曲结构的线状要素Morphing变换方法。该方法首先对不同比例尺表达的对应线状要素分别构建约束Delaunay三角网并建立弯曲森林,然后进行弯曲匹配以获得对应弯曲。鉴于对应弯曲“背面”的独立弯曲结构隐藏于更高层次的大弯曲中,对对应弯曲重新构建约束Delaunay三角网进而建立其“背面”的弯曲森林并进行弯曲匹配得到新的对应弯曲,依此递归充分挖掘对应弯曲结构。在此基础上,将所有对应弯曲的对应始点和对应终点都作为断点切割原线状要素,获得对应线段。最后,采用线性插值算法建立各对应线段之间的对应点关系并以对应点间的直线作为移位路径进行Morphing变换。通过实例分析,验证了本文充分利用独立弯曲结构的方法能够提高对应弯曲特征点的识别能力,从而能够更好地保持弯曲特征点并改善Morphing变换效果。     

利用三维激光扫描技术检测工业构件螺栓孔的空间位置

面向工业检测领域提出了一种工业构件螺栓孔的空间位置检测方法,解决了大型工业钢结构特征位置的检测问题。通过三维激光扫描获得构件的完整点云,提取螺栓孔的点云,利用非线性的最小二乘方法拟合螺栓孔的中心,获得螺栓孔中心在点云坐标系中的坐标。同时,根据设计尺寸建立构件的标准模型作为检测的基准,将模型特征转换成一组点表示,拟合对应的螺栓孔中心,获得螺栓孔中心在模型坐标系中的坐标。利用两个坐标系中若干个公共点计算平移矩阵和旋转矩阵,最后将点云坐标系中的坐标转换到模型坐标系中,计算出标准模型与构件对应点之间的误差,从而实现对工业构件螺栓孔的空间位置精度检测。     

线状特征约束下基于四元数描述的LiDAR点云配准方法

针对经典的基于同名点状特征匹配的LiDAR点云配准算法存在计算量大,点状特征提取精度低,以及基于七参数转换模型的LiDAR点云配准算法中方程线性化过程对配准精度影响较大的特点,提出了以线状特征作为LiDAR点云配准的基元,利用四元数法来表达旋转矩阵,进而形成线状特征约束下基于四元数描述的LiDAR点云配准方法,给出了线状特征约束下三维相似变换的相似性测度表达方法,推导并论证了以线状特征作为配准基元时同名线状特征需要满足的条件。根据四元数与旋转变换矩阵之间的对应关系,求解了基于四元数法的旋转矩阵,并根据旋转矩阵求解了平移及缩放系数。     

位姿信息的小型无人机影像拼接方法

针对传统影像拼接方法需要控制点的问题,该文提出一种基于位姿信息的小型无人机影像拼接方法。首先从机载GPS和惯性导航单元获取无人机拍摄的位置姿态参数,利用Harris算法提取特征点,采用SIFT描述方法计算特征向量,通过近似最近邻算法和PROSAC算法对匹配的同名点对精化。然后根据同名点对像平面坐标和航拍时的位置姿态信息,计算影像之间的变换矩阵,通过最小二乘原理平差计算以提高变换矩阵精度。最后利用变换矩阵实现无人机多幅影像的拼接,得到兴趣区域全景图。实验结果验证了该方法在拼接效果方面具有优势。     

场景判断增强SURF影像匹配算法

SURF特征匹配算法进行特征匹配操作时,会受到边缘线上大量相似点特征等因素的影响,特征匹配结果往往不会达到预期的效果,在纹理特征复杂的场景区域,其匹配精度会变得更低。为了实现在复杂场景中连续影像帧之间高质量的特征匹配,本文提出了一种基于场景判断的改进SURF影像匹配算法。该方法以GC-SURF特征匹配算法为基础,在初始匹配与几何对应匹配阶段增加了变换矩阵评估机制。首先,采用本文算法构建两帧影像间的变换关系模型,求解基础矩阵与单应矩阵;然后,通过变换矩阵评估策略,对上一步求得的基础矩阵与单应矩阵进行评估,判断影像场景的立体程度,根据场景立体程度选择相应的变换矩阵;最后,根据选择的变换矩阵进行后续几何对应匹配。对ICL-NUIM数据的综合试验表明,本文算法能够取得更高质量的匹配效果。     

几何刚性和法向量采样一致性的点云配准算法

针对三维激光点云配准中随机采样一致性(RANSAC)算法存在采样次数多、准确度低的缺点,该文提出了一种结合几何刚性和法向量一致性的点云配准算法。该算法通过改进采样策略降低采样次数并提升匹配精度,首先将点云的法向量信息加入采样集,使得每次的采样点从三对减少为两对;接着以两对采样点的刚性和法向量一致性计算置信度来确定当前采样是否置信;最后以迭代运算选取采样内点数最高的样本来估算变换矩阵实现点云精确配准。对激光三维点云进行配准试验,结果表明,本文方法在匹配效率及匹配性能上均优于传统RANSAC算法,且配准精度更高。     

基于位姿信息的小型无人机影像拼接方法

针对传统影像拼接方法需要控制点的问题,该文提出一种基于位姿信息的小型无人机影像拼接方法。首先从机载GPS和惯性导航单元获取无人机拍摄的位置姿态参数,利用Harris算法提取特征点,采用SIFT描述方法计算特征向量,通过近似最近邻算法和PROSAC算法对匹配的同名点对精化。然后根据同名点对像平面坐标和航拍时的位置姿态信息,计算影像之间的变换矩阵,通过最小二乘原理平差计算以提高变换矩阵精度。最后利用变换矩阵实现无人机多幅影像的拼接,得到兴趣区域全景图。实验结果验证了该方法在拼接效果方面具有优势。     

一种基于激光测距仪的扫描匹配方法

扫描匹配是移动机器人定位和环境构图的关键问题之一,通过计算使相邻扫描重叠度最大的最优变换,从而估算相邻时刻的运动量。ICP算法是目前应用最广泛的扫描匹配算法,针对其对离群扫描点和稀疏扫描点敏感的问题,文中提出一种基于连通格序列的方法对扫描数据进行预处理,从而为扫描匹配提供较为理想的输入条件。此外,针对对应关系中一对多和对应距离极端大的问题,文中通过建立唯一对应,并基于四分位数法计算对应距离的上截断点进行剔除,从而避免建立一对多对应,并减少距离极端大的对应。实验结果表明,文中方法能够剔除扫描中的离群点与稀疏点,并改善ICP算法在实际应用中的扫描匹配表现。     

基于SIFT的宽基线立体影像密集匹配

提出基于对极几何和单应映射双重约束及SIFT特征的宽基线立体影像多阶段准密集匹配算法。算法包括三个阶段：①基于特征点的空间分布和信息熵选取一定数量的最优SIFT特征点集并进行最小二乘初始稀疏匹配及立体像对的基本矩阵和单应矩阵估计;②对于其余特征,利用同名核线倾斜角及SIFT特征的尺度信息对匹配窗口的仿射变换参数进行迭代优化及变形改正、提取仿射不变SIFT特征描述符,并基于双重约束信息及欧氏距离测度进行匹配;③考虑宽基线立体影像较低的特征提取重复率,对第②步左右影像中未能成功匹配的特征点,基于双向搜索策略,采用基于盒滤波加速计算的SSD测度在变形改正后的双重约束区域中进行匹配,并对匹配结果进行加权最小二乘拟合定位。实际的宽基线立体影像试验结果证明了算法的有效性,可为后续的三维重建提供较为可靠的密集或准密集匹配点。     

Multi-view dense matching supported by triangular meshes

We present a new procedure to compute dense 3D point clouds from a sequential set of images. This procedure is considered as a second step of a three-step algorithm for 3D reconstruction from image sequences, whose first step consists of image orientation and the last step is shape reconstruction. We assume that the camera matrices as well as a sparse set of 3D points are available and we strive for obtaining a dense and reliable 3D point cloud. Three novel ideas are presented: (1) for sparse tracking and triangulation, the search space for correspondences is reduced to a line segment by means of known camera matrices and disparity ranges are provided by triangular meshes from the already available points; (2) triangular meshes from extended sets of points are used for dense matching, because these meshes help to reconstruct points in weakly textured areas and present a natural way to obtain subpixel accuracy; (3) two non-local optimization methods, namely, 1D dynamic programming along horizontal lines and semi-global optimization were employed for refinement of local results obtained from an arbitrary number of images. All methods were extensively tested on a benchmark data set and an infrared video sequence. Both visual and quantitative results demonstrate the effectiveness of our algorithm.     

An algorithm for object reconstruction without interior orientation

This paper presents a novel algorithm for object reconstruction without interior orientation. First we introduce the Thompson and Longuet-Higgins equation as well as the fundamental matrix. After defining the affine model, we show that some of its components can be linearly withdrawn from the findamental matrix, which in turn is linearly determined up to a scale factor by minimal eight image correspondences. The object reconstruction is then performed linearly via a 3D affine transformation. Unlike the well-known DLT algorithm where minimum six known points are required on each image of a stereopair, our algorithm requires that only four of them appear on the second image. In addition to an accuracy fully compatible with the DLT algorithm, tests with an aerial stereopair also show the robustness of our algorithm, both to the configuration of known points and to the image deformation.     

A criterion matrix for the second order design of control networks

This paper proposes an original method of constructing a criterion matrix for the optimal design of control networks by means of the contraction of the eigenvalues and the rotation of the eigenvectors of a covariance matrix. The rotation of the eigenvectors is carried out by two different methods, that is through independent rotations of the pairs of the eigenvector matrix components and through a procrustean transfomation of the same matrix. The required criterion is that the “essential eigenvector components” must be as orthogonal as possible to the predicted direction of deformation. A Second Order Design problem is then resolved, that is the optimization of the precision of the observations of a local free distance network to be constructed for the study of recent crustal movements in the seismogenetic area of Friuli (Italy). The results obtained show a high degree of correspondence for the two proposed methods of rotation if limited rotations of the eigenvectors are considered.     

Complex correspondences in straight line stereo matching

Partial occlusion and fragmented lines will result in the various cases of straight line correspondences, such as one-to-one, one-to-many or many-to-many ones. However, the complex correspondences, such as one-to-many and many-to-many ones, are usually ignored or cannot be established completely in the existing methods. Here, the essence of the complex correspondences will be analyzed. Based on the two characteristics of a straight line, which are introduced by regarding a straight line as a set of collinear points, the compatibility between the complex correspondences and the uniqueness constraint of point correspondence is proved and a new uniqueness constraint of correspondence for matching lines is proposed. Based on the analysis of the complex correspondences, a new concept of line feature group is defined to describe a set of integral correspondences among straight line features from different images and then a new algorithm for establishing all the correspondences completely is described simply. The experimental results with real stereo images illustrate that the complex correspondences among straight lines are actual cases and can be established effectively.     

适应任意形状铁路隧道LiDAR点云的断面差异分析方法

针对三维激光扫描技术在铁路隧道超欠挖分析和收敛分析等应用中如何能对任意形状隧道LiDAR点云进行断面差异分析的关键问题,本文提出了一种适应任意形状铁路隧道LiDAR点云的断面差异分析方法。该方法先建立铁路隧道设计断面统一模型,提出实测断面点与设计要素对应关系判定方法,对从隧道LiDAR点云中提取出的断面在铁路隧道设计断面统一模型中逐点确定所对应的设计要素,并计算出断面在对应设计要素上的径向投影点到设定的起算高程面与设计断面左侧交点的弧长,根据弧长对断面进行排序,最后根据设定的采样步长对有序的断面进行采样得到实测断面,进而计算出实测断面相较于设计断面的偏差量。通过3组具有代表性的不同断面隧道模拟点云数据对提出的算法参数最佳取值进行试验,试验结果不仅给出了算法参数的最佳取值,也论证了本方法可以对任意形状铁路隧道LiDAR点云进行断面差异分析。     

Surface Matching and Change Detection Using a Modified Hough Transformation for Robust Parameter Estimation

Many photogrammetric and GIS applications, such as city modelling, change detection and object recognition, deal with surfaces. Change detection involves looking for differences between two surface models that are obtained from different sensors, for example an optical sensor and a laser scanner, or by the same sensor at different epochs. Surfaces obtained through a sampling process may also have to be compared for future processing (for example transformation parameter estimation and change detection). Surface matching is therefore an essential task in these applications. The matching of surfaces involves two steps. The first step deals with finding the correspondences between two surface points and/or patches. The second step requires the determination of transformation parameters between the two surfaces. However, since most surfaces consist of randomly distributed discrete points and may have different reference systems, finding the correspondences cannot be achieved without knowing the transformation parameters between the two surfaces. Conversely, deriving the transformation parameters requires the knowledge of the correspondence between the two point sets. The suggested approach for surface matching deals with randomly distributed data sets without the need for error prone interpolation and requires no point-to-point correspondence between the two surfaces under consideration. This research simultaneously solves for the correspondence and the transformation parameters using a Modified Iterated Hough Transform for robust parameter estimation. Several experiments are conducted to prove the feasibility and the robustness of the suggested approach, even when a high percentage of change exists.     

城市扩展强度及其地表热特性遥感定量分析

城市化将自然景观转变为不同土地利用的城市景观,改变了地表的物理特性。因此,有关城市扩张及其对环境的影响的准确信息具有广泛的应用价值。利用遥感技术监测土地利用变化成为城市环境研究的重要内容,然而利用遥感技术定量提取城市扩展及其强度的定量信息仍存在许多挑战。本文提出了定量确定城市扩展范围及其发展强度的新方法——地图密度指数。并以中部城市——长沙为例,利用Landsat TM/ETM+数据定量评价了城市扩张及其热环境特征的变化。首先,集成遥感光谱指数提取地表非渗透表面,然后利用移动窗口算法获得地图密度指数,再根据设定的阈值获得密度指数等级图,依此密度指数等级图识别城市扩展范围及其发展强度。再结合地表温度反演的数据,分析了城市格局及其变化与地表热特性变化的定量关系。结果表明,自1990年代以来,长沙市城市区域及其发展密度显著增加,城市发展的密度差异与地表温度相一致。 地图密度指数能较好刻画城市扩展范围及其发展强度,并与地表温度空间分布存在较好的对应关系。