地面激光扫描点云与鱼眼全景影像的配准与整合研究
发布时间：2018-10-04 22:13
【摘要】：多传感器信息集成与多源异构数据整合是当前遥感技术的重要发展方向。整合地面激光扫描点云和鱼眼全景影像对于大视域环境感知和理解、虚拟场景分析工作等有重要意义。本文提出了一种面向数据层的激光点云与鱼眼全景影像的数据整合方法：利用柱面模型进行透视影像变换，实现参考系统一；通过直线特征提取与配准实现两种数据基于视点约束下的三维对齐；通过创建虚拟点，采用混合彩色离散点和带地理参考全景影像两种数据形式表达场景信息，最后通过扩展视点类、空间模型类元数据实现二者数据级的整合。论文进行了地面激光点云与鱼眼影像的整合实验，，实验表明论文提出的方法有能力整合多源数据，并为大规模、多场景的数据集成、建库以及跨平台调用提供了数据条件。
[Abstract]:Multi-sensor information integration and multi-source heterogeneous data integration is an important development direction of remote sensing technology. The integration of ground laser scanning point cloud and fish eye panoramic images is of great significance for the perception and understanding of large visual environment and virtual scene analysis. In this paper, a data integration method of laser point cloud and fish-eye panoramic image for data layer is proposed: the perspective image is transformed by cylindrical model, and the reference system is realized. Through line feature extraction and registration, the two kinds of data can be aligned based on viewpoint constraints. By creating virtual points, the scene information can be expressed in the form of mixed color discrete points and panoramic images with geographical reference. Finally, the integration of the two data levels is realized by extending view class and spatial model class metadata. The experiment of integrating ground laser point cloud and fish-eye image shows that the proposed method can integrate multi-source data and provide the data conditions for large-scale, multi-scene data integration, database building and cross-platform call.
【学位授予单位】：中国矿业大学（北京）
【学位级别】：博士
【学位授予年份】：2014
【分类号】：P237;P225.2
本文编号：2252075
[Abstract]:Multi-sensor information integration and multi-source heterogeneous data integration is an important development direction of remote sensing technology. The integration of ground laser scanning point cloud and fish eye panoramic images is of great significance for the perception and understanding of large visual environment and virtual scene analysis. In this paper, a data integration method of laser point cloud and fish-eye panoramic image for data layer is proposed: the perspective image is transformed by cylindrical model, and the reference system is realized. Through line feature extraction and registration, the two kinds of data can be aligned based on viewpoint constraints. By creating virtual points, the scene information can be expressed in the form of mixed color discrete points and panoramic images with geographical reference. Finally, the integration of the two data levels is realized by extending view class and spatial model class metadata. The experiment of integrating ground laser point cloud and fish-eye image shows that the proposed method can integrate multi-source data and provide the data conditions for large-scale, multi-scene data integration, database building and cross-platform call.
【学位授予单位】：中国矿业大学（北京）
【学位级别】：博士
【学位授予年份】：2014
【分类号】：P237;P225.2
【参考文献】
相关期刊论文 前10条
1 张毅;闫利;杨红;陈向阳;;地面三维激光扫描的系统误差模型研究[J];测绘通报;2012年01期
2 张帆;黄先锋;李德仁;;基于球面投影的单站地面激光扫描点云构网方法[J];测绘学报;2009年01期
3 钟成;;集成LiDAR和遥感影像生成高质量真正射影像研究[J];测绘学报;2011年01期
4 王晏民;胡春梅;;一种地面激光雷达点云与纹理影像稳健配准方法[J];测绘学报;2012年02期
5 王敬贵;杜云艳;苏奋振;周成虎;;基于地理本体的空间数据集成方法及其实现[J];地理研究;2009年03期
6 马立广;地面三维激光扫描仪的分类与应用[J];地理空间信息;2005年03期
7 张爱武,胡少兴,孙卫东,李风亭;基于激光与可见光同步数据的室外场景三维重建[J];电子学报;2005年05期
8 张诚,汪嘉业;利用鱼眼照片恢复景物的深度信息[J];工程图学学报;2002年03期
9 杨梦宁;廖训佚;林晓泽;张小洪;;自动修正拍摄俯仰误差的鱼眼图像校正方法[J];光电工程;2010年06期
10 张宁;刘天键;;鱼眼图像校正算法研究与实现[J];闽江学院学报;2012年05期
相关博士学位论文 前1条
1 张志超;融合机载与地面LIDAR数据的建筑物三维重建研究[D];武汉大学;2010年
本文编号：2252075
本文链接：https://www.wllwen.com/kejilunwen/dizhicehuilunwen/2252075.html
