正射影像纠正及拼接线网络自动生成方法研究
发布时间：2018-11-19 07:54
【摘要】：正射影像的无缝拼接技术,是数字正射影像产品(DOM)生产必不可少的一步,无缝自动拼接技术是摄影测量与遥感发展的一个愿景,也对最终DOM产品质量起着关键性的作用。当前,在以航空影像及无人机影像为数据源进行数字正射影像产品生产过程中,影像拼接中的拼接线网络自动优化仍然是制约数字正射影像生产自动化的瓶颈问题。本文以解决航空影像及无人机影像的正射影像产品生产过程中的这个瓶颈问题为目的,对涉及到的若干关键问题进行了研究。主要的研究工作如下： (1)正射影像纠正。针对正射影像计算密集型特点,提出了一种基于OpenMP的影像正射纠正方法,对影像正射纠正进行并行加速,提高程序运行效率,缩短程序运行时间。对多核计算机的计算效率进行了分析,实验结果表明本文方法可以有效保证正射影像纠正的效率。 (2)拼接线网络自动生成。依据一种新的顾及重叠的面Voronoi图,并对基于顾及重叠的面Voronoi图提出了一种接缝线网络的自动生成的方法。解决了重叠区域是凸多边形或凹多边形的中轴线问题,解决了重叠区域交点个数两个以上的情况。并提高了大范围影像镶嵌时的质量和效率问题。本文在生成拼接线网络时,首先是对存在重叠区正射影像面集区域的进行重新划分,根据每张正射影像图所属的有效镶嵌多边形进行扫描线填充来形成最终的影像镶嵌图。生成的拼接线网络是基于全局的,整体考虑了拼接范围内的所有要处理的正射影像,将各个正射影像的覆盖的有效范围划分成了有效的镶嵌多边形,并且这种划分是唯一的、无缝的和无冗余的。 本文提出的正射影像纠正与接缝线网络自动生成方法采用Visual C++进行了实现,并且开发了原型系统,采用实际的无人机影像和航空倾斜影像以及对应的DEM数据进行了实验,结果表明本文方法具有可行性。图29个,表6个,参考文献37个
[Abstract]:Seamless mosaic technology of orthophoto image is an indispensable step in the production of digital orthophoto product (DOM). Seamless automatic stitching technology is a vision of photogrammetry and remote sensing development, and also plays a key role in the final product quality of DOM. At present, in the production process of digital orthophoto products based on aerial images and UAV images, the automatic optimization of stitching line in image mosaic is still a bottleneck problem that restricts the automation of digital orthophoto production. In order to solve the bottleneck problem in the production of aerial image and UAV image orthophoto products, this paper studies some key problems involved. The main research work is as follows: (1) orthophoto correction. In view of the computational intensive characteristics of orthophoto, a method of orthophoto correction based on OpenMP is proposed, which accelerates the correction of orthophoto in parallel, improves the efficiency of the program and shortens the running time of the program. The computational efficiency of multi-core computer is analyzed. The experimental results show that the proposed method can effectively guarantee the efficiency of orthophoto correction. (2) automatic generation of splicing network. Based on a new surface Voronoi graph with overlapping in mind, an automatic generation method of seam line network based on the overlapping plane Voronoi graph is proposed. The problem that overlapping region is the central axis of convex polygon or concave polygon is solved, and more than two intersection points of overlapping region are solved. It also improves the quality and efficiency of large-scale image mosaic. In this paper, the first step is to redivide the area of orthophoto surface set with overlapped area, and fill the scanning line according to the effective mosaic polygon of each orthophoto image to form the final image mosaic image. The generated splicing network is based on the whole situation. It takes into account all the orthophoto images to be processed in the mosaic range, and divides the effective range of each orthophoto image into effective mosaic polygons, and this partition is unique. Seamless and non-redundant. In this paper, the orthophoto correction and seam line network automatic generation method is implemented by Visual C, and a prototype system is developed. The actual UAV image, the aerial tilt image and the corresponding DEM data are used to experiment. The results show that this method is feasible. 29 figures, 6 tables, 37 references
【学位授予单位】：中南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：P231
本文编号：2341614
[Abstract]:Seamless mosaic technology of orthophoto image is an indispensable step in the production of digital orthophoto product (DOM). Seamless automatic stitching technology is a vision of photogrammetry and remote sensing development, and also plays a key role in the final product quality of DOM. At present, in the production process of digital orthophoto products based on aerial images and UAV images, the automatic optimization of stitching line in image mosaic is still a bottleneck problem that restricts the automation of digital orthophoto production. In order to solve the bottleneck problem in the production of aerial image and UAV image orthophoto products, this paper studies some key problems involved. The main research work is as follows: (1) orthophoto correction. In view of the computational intensive characteristics of orthophoto, a method of orthophoto correction based on OpenMP is proposed, which accelerates the correction of orthophoto in parallel, improves the efficiency of the program and shortens the running time of the program. The computational efficiency of multi-core computer is analyzed. The experimental results show that the proposed method can effectively guarantee the efficiency of orthophoto correction. (2) automatic generation of splicing network. Based on a new surface Voronoi graph with overlapping in mind, an automatic generation method of seam line network based on the overlapping plane Voronoi graph is proposed. The problem that overlapping region is the central axis of convex polygon or concave polygon is solved, and more than two intersection points of overlapping region are solved. It also improves the quality and efficiency of large-scale image mosaic. In this paper, the first step is to redivide the area of orthophoto surface set with overlapped area, and fill the scanning line according to the effective mosaic polygon of each orthophoto image to form the final image mosaic image. The generated splicing network is based on the whole situation. It takes into account all the orthophoto images to be processed in the mosaic range, and divides the effective range of each orthophoto image into effective mosaic polygons, and this partition is unique. Seamless and non-redundant. In this paper, the orthophoto correction and seam line network automatic generation method is implemented by Visual C, and a prototype system is developed. The actual UAV image, the aerial tilt image and the corresponding DEM data are used to experiment. The results show that this method is feasible. 29 figures, 6 tables, 37 references
【学位授予单位】：中南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：P231
【参考文献】
相关期刊论文 前10条
1 陈冉丽;吴侃;;三维激光扫描用于获取开采沉陷盆地研究[J];测绘工程;2012年03期
2 潘俊;王密;李德仁;;接缝线网络的自动生成及优化方法[J];测绘学报;2010年03期
3 陈文通;大比例尺彩色数字正射影像地图制作及建库技术初探[J];城市勘测;2005年04期
4 喜文飞;方源敏;李帅;李健;;一种新的激光点云数据精简方法[J];测绘工程;2012年04期
5 苟卫涛;翟群英;戚娅萍;李英利;段英;;基于JX4C DPW实现IRS-P5卫星影像定向建模方法探讨[J];测绘标准化;2013年01期
6 张剑清;柯涛;孙明伟;;基于集群计算机的海量航空数码影像并行处理——并行计算在航空数字摄影测量中的应用[J];计算机工程与应用;2008年13期
7 吴壮志;;关于计算机系开设“计算几何”课程的建议[J];计算机教育;2009年21期
8 王浩;姚一飞;;正射影像地图的制作及其应用[J];科技致富向导;2011年05期
9 刘勇奎,高云,黄有群;一个有效的多边形裁剪算法[J];软件学报;2003年04期
10 王妍颖;马艳;刘定生;;一种基于动态分组策略和多线程并行IO的并行镶嵌算法优化[J];遥感信息;2012年02期
，本文编号：2341614
本文链接：https://www.wllwen.com/kejilunwen/dizhicehuilunwen/2341614.html
