无人机遥感影像与数字高程模型的三维可视化研究
发布时间：2018-09-08 13:47
【摘要】：在近些年来,随着遥感,GPS导航以及计算机等技术领域的发展,多技术集成的无人机遥感技术顺势而起。其应用领域也不断拓展,在动态监测,数字化建设以及灾害应急等多领域有着成熟的应用。对于无人机这种尺度变化及旋转角都大的影像,关于其拼接匹配技术的研究也比较多。但是,针对大比例尺无人机影像提取高分辨率DEM的研究相对较少,如何根据影像地形特征快速建立数字可视化模型是无人机遥感影像应用成果的体现,也是其深入各领域发展的前提。文章首先概述了当前无人机影像在三维制作上的应用研究现状。针对影像分辨率高,且具有一定重叠度等特点,提出了基于无人机影像快速建立三维数字化模型的方法。 研究工作包括以下几个方面： (1)针对当前大比例尺无人机影像制作DEM数据的方法研究中存在精度不高以及缺乏有效的质量评定等问题,文章引入了中误差及地面粗糙度等评价因子对利用地形特征点提取高分辨率DEM数据进行了质量评价,从而避免了DEM数据在地形数字化应用方面的盲目性。 (2)考虑无人机影像地形高程差的影响,从引起影像变形误差的原理出发,分析了无人机影像几何变形的过程,并且主要对由外方位元素引起的误差进行了纠正处理。利用特征控制点并结合多项式及共线方程方法对不同地形加以处理,使得后期影像数据在三维应用上的结果更加理想。 (3)针对大数据量,尤其是对于大规模城市区域建立数字化模型时,建模速度慢,显示效率不高。本文结合传统分块算法和三角网生成算法,提出了一种自适用阈值分块构网算法,从而提高了建模速度,在保证一定质量的条件下也提高了自动化处理水平。 (4)结合传统二叉树和四叉树思想建立了一种基于观察点的多分辨率模型,根据地形特征点与观察点的距离大小,分析其分辨率的动态变换关系,根据地形复杂度合适采用细化、简化处理,有效解决了三维可视化模型显示效率不高的问题。应用该模型分别在百度地图与Google earth上实现了影像二维叠加以及三维航迹可视化模拟。图48幅,表4个,参考文献52篇。
[Abstract]:In recent years, with the development of GPS navigation and computer technology, multi-technology integrated remote sensing technology of UAV has come into being. Its application field also expands unceasingly, has the mature application in the dynamic monitoring, the digitization construction and the disaster emergency and so on many fields. For UAV images with large scale variation and rotation angle, there are more researches on the matching techniques. However, there is relatively little research on high-resolution DEM extraction from large scale UAV images. How to quickly establish a digital visualization model based on the terrain features of UAV images is the embodiment of UAV remote sensing image application results. It is also the premise of its in-depth development in various fields. Firstly, the paper summarizes the application and research status of UAV image in three-dimensional production. Aiming at the characteristics of high resolution and certain overlap, a method of building 3D digital model based on UAV image is proposed. The research work includes the following aspects: (1) there are some problems such as low precision and lack of effective quality evaluation in the research of DEM data for large scale UAV images. In this paper, the evaluation factors such as middle error and surface roughness are introduced to evaluate the quality of high resolution DEM data extracted from terrain feature points. Therefore, the blindness of DEM data in terrain digitization is avoided. (2) considering the influence of terrain elevation difference of UAV image, the process of geometric deformation of UAV image is analyzed based on the principle of causing image deformation error. The error caused by the external azimuth element is corrected. By using characteristic control points, polynomial and collinear equation method to deal with different terrain, the result of 3D application of late image data is more ideal. (3) aiming at the large amount of data, Especially for large-scale urban areas, the modeling speed is slow and the display efficiency is not high. In this paper, combining with traditional block algorithm and triangular mesh generation algorithm, a self-adaptive threshold partition algorithm is proposed, which improves the speed of modeling. The automatic processing level is also improved under the condition of certain quality. (4) combined with the traditional binary tree and quadtree idea, a multi-resolution model based on observation point is established, according to the distance between terrain feature point and observation point, The dynamic transformation relation of its resolution is analyzed. According to the terrain complexity, thinning and simplification are adopted, which can effectively solve the problem that the display efficiency of 3D visualization model is not high. The model is used to simulate the two dimensional superposition of image and the visualization of 3D track on Baidu map and Google earth respectively. 48 figures, 4 tables, 52 references.
【学位授予单位】：中南大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P208;P231
本文编号：2230699
[Abstract]:In recent years, with the development of GPS navigation and computer technology, multi-technology integrated remote sensing technology of UAV has come into being. Its application field also expands unceasingly, has the mature application in the dynamic monitoring, the digitization construction and the disaster emergency and so on many fields. For UAV images with large scale variation and rotation angle, there are more researches on the matching techniques. However, there is relatively little research on high-resolution DEM extraction from large scale UAV images. How to quickly establish a digital visualization model based on the terrain features of UAV images is the embodiment of UAV remote sensing image application results. It is also the premise of its in-depth development in various fields. Firstly, the paper summarizes the application and research status of UAV image in three-dimensional production. Aiming at the characteristics of high resolution and certain overlap, a method of building 3D digital model based on UAV image is proposed. The research work includes the following aspects: (1) there are some problems such as low precision and lack of effective quality evaluation in the research of DEM data for large scale UAV images. In this paper, the evaluation factors such as middle error and surface roughness are introduced to evaluate the quality of high resolution DEM data extracted from terrain feature points. Therefore, the blindness of DEM data in terrain digitization is avoided. (2) considering the influence of terrain elevation difference of UAV image, the process of geometric deformation of UAV image is analyzed based on the principle of causing image deformation error. The error caused by the external azimuth element is corrected. By using characteristic control points, polynomial and collinear equation method to deal with different terrain, the result of 3D application of late image data is more ideal. (3) aiming at the large amount of data, Especially for large-scale urban areas, the modeling speed is slow and the display efficiency is not high. In this paper, combining with traditional block algorithm and triangular mesh generation algorithm, a self-adaptive threshold partition algorithm is proposed, which improves the speed of modeling. The automatic processing level is also improved under the condition of certain quality. (4) combined with the traditional binary tree and quadtree idea, a multi-resolution model based on observation point is established, according to the distance between terrain feature point and observation point, The dynamic transformation relation of its resolution is analyzed. According to the terrain complexity, thinning and simplification are adopted, which can effectively solve the problem that the display efficiency of 3D visualization model is not high. The model is used to simulate the two dimensional superposition of image and the visualization of 3D track on Baidu map and Google earth respectively. 48 figures, 4 tables, 52 references.
【学位授予单位】：中南大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P208;P231
【参考文献】
相关期刊论文 前10条
1 谷宁;冯仲科;罗旭;;基于无人飞行器航测建立数字高程模型研究[J];北京林业大学学报;2007年S2期
2 李军;李永树;蔡国林;;利用无人机影像制作地震灾区三维景观图[J];测绘工程;2012年01期
3 崔红霞,林宗坚,孙杰;大重叠度无人机遥感影像的三维建模方法研究[J];测绘科学;2005年02期
4 范承啸;韩俊;熊志军;赵毅;;无人机遥感技术现状与应用[J];测绘科学;2009年05期
5 陈飞;吴英男;;基于数字高程模型和遥感影像的三维可视化[J];测绘科学;2009年S2期
6 胡荣明;张敏;刘潘;;高分辨率无人机低空影像DEM的建立及其精度研究[J];测绘科学;2011年04期
7 张丽丽;王小平;张瑛;;基于无人机影像生产高精度DEM的实践[J];测绘技术装备;2009年01期
8 崔红霞,林宗坚,孙杰;无人机遥感监测系统研究[J];测绘通报;2005年05期
9 刘异;李玉霞;童玲;;无地面控制点的无人机遥感影像几何校正算法[J];测绘通报;2012年07期
10 卜彦龙;潘亮;彭辉;沈林成;;未知DEM下的机载SAR图像几何校正方法研究[J];测绘学报;2009年01期
，本文编号：2230699
本文链接：https://www.wllwen.com/kejilunwen/dizhicehuilunwen/2230699.html
