無人機遙感影像與數(shù)字高程模型的三維可視化研究
發(fā)布時間：2018-09-08 13:47
【摘要】：在近些年來,隨著遙感,GPS導航以及計算機等技術領域的發(fā)展,多技術集成的無人機遙感技術順勢而起。其應用領域也不斷拓展,在動態(tài)監(jiān)測,數(shù)字化建設以及災害應急等多領域有著成熟的應用。對于無人機這種尺度變化及旋轉角都大的影像,關于其拼接匹配技術的研究也比較多。但是,針對大比例尺無人機影像提取高分辨率DEM的研究相對較少,如何根據(jù)影像地形特征快速建立數(shù)字可視化模型是無人機遙感影像應用成果的體現(xiàn),也是其深入各領域發(fā)展的前提。文章首先概述了當前無人機影像在三維制作上的應用研究現(xiàn)狀。針對影像分辨率高,且具有一定重疊度等特點,提出了基于無人機影像快速建立三維數(shù)字化模型的方法。 研究工作包括以下幾個方面： (1)針對當前大比例尺無人機影像制作DEM數(shù)據(jù)的方法研究中存在精度不高以及缺乏有效的質量評定等問題,文章引入了中誤差及地面粗糙度等評價因子對利用地形特征點提取高分辨率DEM數(shù)據(jù)進行了質量評價,從而避免了DEM數(shù)據(jù)在地形數(shù)字化應用方面的盲目性。 (2)考慮無人機影像地形高程差的影響,從引起影像變形誤差的原理出發(fā),分析了無人機影像幾何變形的過程,并且主要對由外方位元素引起的誤差進行了糾正處理。利用特征控制點并結合多項式及共線方程方法對不同地形加以處理,使得后期影像數(shù)據(jù)在三維應用上的結果更加理想。 (3)針對大數(shù)據(jù)量,尤其是對于大規(guī)模城市區(qū)域建立數(shù)字化模型時,建模速度慢,顯示效率不高。本文結合傳統(tǒng)分塊算法和三角網(wǎng)生成算法,提出了一種自適用閾值分塊構網(wǎng)算法,從而提高了建模速度,在保證一定質量的條件下也提高了自動化處理水平。 (4)結合傳統(tǒng)二叉樹和四叉樹思想建立了一種基于觀察點的多分辨率模型,根據(jù)地形特征點與觀察點的距離大小,分析其分辨率的動態(tài)變換關系,根據(jù)地形復雜度合適采用細化、簡化處理,有效解決了三維可視化模型顯示效率不高的問題。應用該模型分別在百度地圖與Google earth上實現(xiàn)了影像二維疊加以及三維航跡可視化模擬。圖48幅,表4個,參考文獻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.
【學位授予單位】：中南大學
【學位級別】：碩士
【學位授予年份】：2013
【分類號】：P208;P231
本文編號：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.
【學位授予單位】：中南大學
【學位級別】：碩士
【學位授予年份】：2013
【分類號】：P208;P231
【參考文獻】
相關期刊論文 前10條
1 谷寧;馮仲科;羅旭;;基于無人飛行器航測建立數(shù)字高程模型研究[J];北京林業(yè)大學學報;2007年S2期
2 李軍;李永樹;蔡國林;;利用無人機影像制作地震災區(qū)三維景觀圖[J];測繪工程;2012年01期
3 崔紅霞,林宗堅,孫杰;大重疊度無人機遙感影像的三維建模方法研究[J];測繪科學;2005年02期
4 范承嘯;韓俊;熊志軍;趙毅;;無人機遙感技術現(xiàn)狀與應用[J];測繪科學;2009年05期
5 陳飛;吳英男;;基于數(shù)字高程模型和遙感影像的三維可視化[J];測繪科學;2009年S2期
6 胡榮明;張敏;劉潘;;高分辨率無人機低空影像DEM的建立及其精度研究[J];測繪科學;2011年04期
7 張麗麗;王小平;張瑛;;基于無人機影像生產(chǎn)高精度DEM的實踐[J];測繪技術裝備;2009年01期
8 崔紅霞,林宗堅,孫杰;無人機遙感監(jiān)測系統(tǒng)研究[J];測繪通報;2005年05期
9 劉異;李玉霞;童玲;;無地面控制點的無人機遙感影像幾何校正算法[J];測繪通報;2012年07期
10 卜彥龍;潘亮;彭輝;沈林成;;未知DEM下的機載SAR圖像幾何校正方法研究[J];測繪學報;2009年01期
，本文編號：2230699
本文鏈接：http://sikaile.net/kejilunwen/dizhicehuilunwen/2230699.html
最近更新
教材專著
