共查询到20条相似文献，搜索用时 125 毫秒
1.
一种基于无人机序列图像的地形地貌三维快速重建方法 总被引：1，自引：0，他引：1
提出了一种基于无人机序列图像的地形地貌三维重建方法,该方法采用Harris特征点和SIFT特征向量来提取图像特征,实现图像配准;采用准透视投影模型和因子化方法对未标定的图像序列进行自动标定;通过高效次优解三角化方法获取三维点云坐标;通过准稠密化扩散算法对三维点云进行稠密化;采用捆绑调整算法提高了空间三维点云的精度;采用Possion表面重建方法对三维点云进行了网格化处理.本文为无人机序列图像的应用提供了一个新的思路,拓展了无人机的应用空间. 相似文献
2.
提出了一种单视影像下的人脸快速三维重建方案,能够对非可控条件下获取的单幅人脸图像进行有效、快速的三维重建。在三维人脸重建阶段,提出了一种特征动态选择策略,针对待重建人脸图像动态选择合适的2D-3D特征变换子空间,在保证执行效率的同时能显著提高三维人脸模型的重建精度。实验证明,本方法不仅简洁高效,而且对人脸姿态和光照变化具较好的鲁棒性,在WHU-3DFaceData数据库上取得了理想的重建效果。 相似文献
3.
倾斜影像之间因存在较大几何畸变,已有算法尚难取得可靠准稠密匹配,严重阻碍后续的三维重建。本文提出一种基于仿射不变特征的自动准稠密匹配与三维重建算法。算法分三个阶段:1)种子匹配。先利用仿射不变特征获取初始匹配,在此基础上进行最小二乘匹配(least square matching,LSM),并自适应迭代求取LSM最大收敛区域(maximum convergence region,MCR);2)准稠密匹配。对位于MCR内部的格网点采用透视变换模型的归一化互相关(normalized cross correlation,NCC)测度进行匹配,其余格网点则采用对极几何改正模型的加权差值平方和测度进行匹配,对于匹配结果中可能存在的外点根据几何一致性关系剔除;3)三维表面重建。利用影像定向与共线方程对2)步结果重建三维点云,然后根据三维点云生成三维网格并贴上纹理,得到最后的重建结果。实际的倾斜立体影像实验验证了算法的优越性。 相似文献
4.
针对传统三维人脸模型构建复杂、效率低下等不足,提出了一种基于单张照片的真实感人脸快速生成方法。首先,构建标准化的人脸模型库,交互式选取输入照片的面部特征点,并基于脸型特征,匹配最佳模型;其次,通过三角形变形和双线性插值的方法实现照片到三维模型的纹理映射,并引入Alpha图实现人脸覆盖纹理到模型中性纹理的融合过渡;最后,采用网格调节方法从整体到细节分层次调整模型以达到生成真实感三维人脸的目的。实验结果证明,该方法对用户友好,所需特征点较少,可实时生成高质量的真实感模型。 相似文献
5.
6.
传统的基于全站仪等单点测量技术难以获取钢结构整体信息,且测量速度慢、模型重建精度低。针对此问题,本文将三维激光扫描技术应用于钢结构模型重建领域,提出了基于LiDAR点云的复杂钢结构模型重建方法。首先,基于标靶球和钢结构面信息实现多测站点云的高精度配准,并通过半径滤波算法去除噪点,以得到完整的激光点云数据;然后,对复杂钢结构单体模型进行分割,以实现各个部件的模型重建;最后,根据现场实测数据对重建的三维模型进行精度评估。试验结果表明,基于该方法得到的模型精度较高,与实际尺寸差值均小于1 cm,为复杂钢结构的三维模型重建提供了有效解决方案。 相似文献
7.
针对基于表面的三维模型形变方法可描述细粒度形变但不独立于三维几何表示,而基于空间的方法则刚好相反这一问题,该文提出了一种基于三维曲面调和场构建的三维形变模型。通过计算表面调和场,将模型划分为若干个形变单元;通过在每个形变单元上运用刚性形状匹配保证局部刚性形变,实现了三维模型的细节保持形变;基于调和值将形变单元上的形变插值到整个三维模型,支持三维模型的细粒度形变;为实现几何表示不相关性,该文提出了一种三维点云调和场计算方法。实验结果表明,该形变模型利用调和场将基于表面和基于空间两种方法的优点结合了起来,具有较好的实时性和可用性。 相似文献
8.
针对增量式SFM算法过渡依赖初始图像的选择,重建时误差累积导致场景漂移且计算效率低的问题,设计了一种改进的加速KAZE特征提取的全局式SFM三维重建算法.首先通过增强特征接受阈值并构造非线性尺度空间来提取图像的特征,使用改进局部差分二进制MLDB描述符对特征进行描述;然后利用全局式SFM算法对整个场景的相机位姿进行估计,从而减小误差并减少光束法平差BA的使用次数,降低了算法的复杂度,大大提高计算效率;最后利用基于面元的多视立体匹配PMVS算法得到稠密重建结果.实验结果表明,同等条件下该方法能够得到更多的有效点对,计算效率约为增量式SFM算法的2倍,且特征匹配准确率提高5％,稠密重建的点云精度高,信息丰富且纹理清晰,是一种有效的三维重建方法. 相似文献
9.
针对三维城市模型重建,提出了一种利用倾斜影像实现城市模型纹理快速重建的方法。该方法基于摄影测量计算机视觉理论,在事先处理得到城市建筑物数字表面模型的基础上,利用共线方程计算物方与像方的几何投影关系,获得建筑物的三维信息与纹理信息。通过一定的最优算法选择出地物面的理想纹理,实现城市建筑物纹理的自动提取及稠密建筑物纹理的遮挡处理。真实影像纹理重建的结果表明:应用本文方法进行三维城市模型纹理重建具有自动化程度高、效果逼真、成本低的特点,为快速、大面积的城市真三维模型纹理重建提供了一种有效的实现途径。 相似文献
10.
待配准的两景影像之间存在非均匀形变时,需要采用非刚性模型进行校正,而密集的、均匀分布的控制点是构建非刚性模型的基础,对此提出一种仿射不变量支持的准稠密控制点匹配算法。首先从两景影像提取特征点并进行稀疏匹配,根据稀疏匹配点消除两景影像间的整体偏移;然后以稀疏匹配点作为三角形的顶点,迭代构建特征点间的近邻三角形,并以面积之比作为仿射不变量,据此判断三角形顶点之间是否满足匹配关系,实现匹配点传播,从而获得准稠密的匹配点集;最后采用HJ-Landsat数据与南极地区Landsat 7 ETM+数据进行实验,结果表明,本文方法能够提取均匀分布的准稠密控制点,在提取率等指标方面优于对比方法。 相似文献
11.
12.
Jinjin Yan Abdoulaye A. Diakit Sisi Zlatanova Mitko Aleksandrov 《Transactions in GIS》2020,24(2):371-389
Outdoor navigation is widely used in daily life, but faces various issues related to the fidelity of outdoor navigation networks. For instance, agents (pedestrians) are often guided via unrealistic detours around places without clear paths (e.g., squares) or if there are vertical constraints such as overpasses/bridges. This is partly explained by the fact that the main sources of navigation networks in current outdoor navigation are two‐dimensional road/street networks. Utilizing a three‐dimensional space‐based navigation model, compatible with some indoor approaches, is a popular way to address the above‐mentioned issues. A 3D space‐based navigation model is generated by treating 3D spaces as nodes and the shared faces as edges. Inputs of this model are enclosed 3D spaces (volumes). However, outdoor spaces are generally open and unbounded. This article puts forward an approach to enclose outdoor spaces and mimic the indoor environments to derive a network based on connectivity and accessibility of spaces. The approach uses 2.5D maps and consists of three major steps: object footprint determination, footprint classification and space creation. Two use cases demonstrate the proposed approach. Enclosing outdoor spaces opens a new research direction toward providing seamless indoor/outdoor navigation for a range of agents. 相似文献
13.
针对三维地质建模工作繁琐、复杂,从手工绘制地质模型角度提出了一种基于GIS耦合SketchUp技术的三维地层模型绘制方法。采用钻孔资料作为数据源,通过ArcGIS的插值和栅格转换工具生成地层三角网格。然后利用SketchUp的封面插件和表面求交工具,生成地层的三维体模型。该方法操作简单、直观和易于掌握,具有一定的推广和应用价值。 相似文献
14.
Shen Ying Renzhong Guo Lin Li Peter Van Oosterom Jantien Stoter 《Transactions in GIS》2015,19(5):758-779
This article presents a new method to illustrate the feasibility of 3D topology creation. We base the 3D construction process on testing real cases of implementation of 3D parcels construction in a 3D cadastral system. With the utilization and development of dense urban space, true 3D geometric volume primitives are needed to represent 3D parcels with the adjacency and incidence relationship. We present an effective straightforward approach to identifying and constructing the valid volumetric cadastral object from the given faces, and build the topological relationships among 3D cadastral objects on‐the‐fly, based on input consisting of loose boundary 3D faces made by surveyors. This is drastically different from most existing methods, which focus on the validation of single volumetric objects after the assumption of the object's creation. Existing methods do not support the needed types of geometry/topology (e.g. non 2‐manifold, singularities) and how to create and maintain valid 3D parcels is still a challenge in practice. We will show that the method does not change the faces themselves and faces in a given input are independently specified. Various volumetric objects, including non‐manifold 3D cadastral objects (legal spaces), can be constructed correctly by this method, as will be shown from the results. 相似文献
15.
融合局部形变模型的鲁棒性人脸识别 总被引：2，自引：1，他引：1
将测试图像看成是人脸库的线性组合,并用形变模型表示,优化匹配求解组合系数,根据系数的稀疏性进行分类识别。为了进一步提高算法的鲁棒性,采用了分片加权的策略。在公用的人脸数据库上进行广泛的实验,结果表明,平均识别率达到97%以上,在遮挡30%时其识别率仍达到95%以上。本方法对人脸识别问题非常有效,且可以显著提高对伪装、遮挡变化的鲁棒性和稳定性。 相似文献
16.
在地质建模中,基于层面的地质块体结构模型的获取是其中最重要的环节之一,目前的构建方法仅关注地质块体的提取和算法效率的提高,忽略了对地质块体与其组成面片以及各地质块体之间拓扑关系的建立,为此,本文提出了一种新的构建方法,基本思路是：首先,利用边界线与曲面片的关系,提取方向边和方向三角形,其次,按照右手法则建立面片排序表,最后,按照一定的遍历规则获取每个面片的正体和负体。此方法不仅可以自动确定地质块体与地质曲面片的组成关系,还可间接分析地质块体之间的邻接关系,从而为三维地质体模型的建立以及三维地质体拓扑查询和分析奠定基础。 相似文献
17.
In the task of 3D building model reconstruction from point clouds we face the problem of recovering a roof topology graph in the presence of noise, small roof faces and low point densities. Errors in roof topology graphs will seriously affect the final modelling results. The aim of this research is to automatically correct these errors. We define the graph correction as a graph-to-graph problem, similar to the spelling correction problem (also called the string-to-string problem). The graph correction is more complex than string correction, as the graphs are 2D while strings are only 1D. We design a strategy based on a dictionary of graph edit operations to automatically identify and correct the errors in the input graph. For each type of error the graph edit dictionary stores a representative erroneous subgraph as well as the corrected version. As an erroneous roof topology graph may contain several errors, a heuristic search is applied to find the optimum sequence of graph edits to correct the errors one by one. The graph edit dictionary can be expanded to include entries needed to cope with errors that were previously not encountered. Experiments show that the dictionary with only fifteen entries already properly corrects one quarter of erroneous graphs in about 4500 buildings, and even half of the erroneous graphs in one test area, achieving as high as a 95% acceptance rate of the reconstructed models. 相似文献
18.
通过对两因子模型的进一步深化、提炼和改进,提出了一种基于两因子模型的多姿态人脸识别方法,该方法能有效地缓解人脸特征对姿态变化较为敏感的问题。实验结果表明,经过姿态因子分离后的人脸全局或局部特征在保持较高显著性的同时,均对姿态变化具有理想的鲁棒性,在FERET人脸数据库上取得了最高92.5%的识别率。 相似文献
19.
基于几何元素的三维景观实体建模研究 总被引：1，自引：0，他引：1
在分析已有国内外研究成果的基础上 ,提出了基于几何元素的 3DCM空间数据模型。实验证明 ,该模型能较精确地表达 3DCM中各类景观实体及其空间关系。 相似文献
20.
Due to the fast development of the urban environment, the need for efficient maintenance and updating of 3D building models is ever increasing. Change detection is an essential step to spot the changed area for data (map/3D models) updating and urban monitoring. Traditional methods based on 2D images are no longer suitable for change detection in building scale, owing to the increased spectral variability of the building roofs and larger perspective distortion of the very high resolution (VHR) imagery. Change detection in 3D is increasingly being investigated using airborne laser scanning data or matched Digital Surface Models (DSM), but rare study has been conducted regarding to change detection on 3D city models with VHR images, which is more informative but meanwhile more complicated. This is due to the fact that the 3D models are abstracted geometric representation of the urban reality, while the VHR images record everything. In this paper, a novel method is proposed to detect changes directly on LOD (Level of Detail) 2 building models with VHR spaceborne stereo images from a different date, with particular focus on addressing the special characteristics of the 3D models. In the first step, the 3D building models are projected onto a raster grid, encoded with building object, terrain object, and planar faces. The DSM is extracted from the stereo imagery by hierarchical semi-global matching (SGM). In the second step, a multi-channel change indicator is extracted between the 3D models and stereo images, considering the inherent geometric consistency (IGC), height difference, and texture similarity for each planar face. Each channel of the indicator is then clustered with the Self-organizing Map (SOM), with “change”, “non-change” and “uncertain change” status labeled through a voting strategy. The “uncertain changes” are then determined with a Markov Random Field (MRF) analysis considering the geometric relationship between faces. In the third step, buildings are extracted combining the multispectral images and the DSM by morphological operators, and the new buildings are determined by excluding the verified unchanged buildings from the second step. Both the synthetic experiment with Worldview-2 stereo imagery and the real experiment with IKONOS stereo imagery are carried out to demonstrate the effectiveness of the proposed method. It is shown that the proposed method can be applied as an effective way to monitoring the building changes, as well as updating 3D models from one epoch to the other. 相似文献