透過您的圖書館登入
IP:44.197.108.187
  • 學位論文

地面控制點對無人飛行載具數值地形模型精度影響之評估

Quality evaluation of unmanned aerial vehicle-associated digital terrain model by means of ground control points

指導教授 : 張國楨

摘要


無人飛行載具(Unmanned Aerial Vehicle, UAV)近年來新興熱門且受到關注,在防災方面擁有許多優勢亦有許多發展空間,非常幸運的本研究室有這方面充足且良好的資源,期望能藉由本研究尋找UAV所拍攝影像的加值作用,對未來災害研究有所貢獻,數值地形模型在地形判識、災害評估方面為最基礎也是最重要的資料,搭配無人載具與控制點能達到何種程度的精度、如何提高精度的最有效率方法為本研究之主要目的。 於市面上無人載具大致分為定翼型與旋翼型兩大類,各有其優缺點,雖說飛機飛行技術發展已經趨於成熟,但風險依然很高;在資源充足情況下對本研究室所擁有的定翼與旋翼兩種無人載具所拍攝影像建置數值地形模型的精度比較,包括搭載相機不同、無人飛行載具不同、飛行特性不同,在同一地區拍攝的成果建置模型進行討論,能在未來的飛行任務針對精度需求選取風險最低且適合的飛行載具,研究成果顯示定翼或旋翼機並不會對精度造成影響,但搭載的相機和拍攝像片品質會影響精度,在像片與模型解析度的要求則與UAV飛行高度有關。 地面控制點(Ground Control Point, GCP)對建置數值地形模型(digital terrain model, DTM)來說有絕對的影響,包括設置點位與數量的設計,無奈礙於現實,可能地形因素、區域衛星訊號受遮蔽或是時間成本關係,我們不可能在飛行區域內布滿控制點,分析地面控制點的數量及分布對精度的影響並進行量化,研究成果顯示控制點分布最為重要,由其最外圍控制點位於研究區域外側一點佳,減少控制點外側誤差較大的影響,並與本研究團隊在現地進行的精密水準高程測量做比較,討論誤差狀況與目前數值地型模型能夠做到怎樣的精確度,並整理出影像要求與精度規範,而目前本研究之高成精度位於±20cm以內,期望能使未來在每一趟飛行任務尋找與設置控制點方面,事先達到最有效率的規劃。

並列摘要


In recent years, Unmanned Aerial Vehicle (UAV) become a popular technology, and is very useful for natural disaster assessment and hazard mitigation study. With good UAV equipment and resources, this study is thus focus on the feasibility and adaptability analysis of the UAV techniques and its’ applications. Digital terrain model (DTM) is the one of the most essential data set for hazard analysis. In this study we try to find what the data and the quality that UAV-associated equipment may produce, and the relationship related with the ground control points (GCPs), and what is the most efficient way to improve the quality, and how to achieve this goals. There are two kinds of UAVs, the fixed wing and the rotor wings, depending on different role of classification. In this study, we compare the quality evaluation of DTM from these two kinds of UAV. Different factors are also evaluated in this study, including different kinds of digital cameras, different UAVs, and different sets of GCP groups, by comparing the DTM quality that applying in the same area. The results indicates that the quality of images affects the DTM, and the quality of image relates with, and only with the mission fly high. However, different UAV platform is not important for data resolution. In this study, we analyzed how does GCP affects DTM, compared with the airborne LiDAR data, and with ground leveling. Even caused of the field situations, time costs… etc. we could not set GCP homogeneously and ideally. The results finds the distribution and the amount of the ground control points are the dominant factors affecting DTM quality. The current result shows that the precision of the DTM could be better than 20cm, compared with airborne LiDAR data. Based on the objective of this study, some suggests and results related with different platforms and equipment selection, and the mission planning is thus discussed.

參考文獻


[14] 施錦揮、游政恭、鄒慶敏、蔡季欣、林志清、林燕山,「無人飛行載具應用於防救災圖資供應之研究—以北二高崩塌地為例」,地籍測量:中華民國地籍測量學會會刊,第29卷,第3期,2010,第17-36頁。
[7] 陳英煥、蔡展榮,「空照數位像機拍攝高重疊影像匹配高密度點雲」,航測及遙測學刊,第十三卷,第三期,2008,第219-230 頁。
[18] 蔡依庭,UAV航拍影像點雲產生DSM之研究,碩士論文,國立台北大學不動產與城鄉環境學系,台北,2012。
[16] 申承翰,無人飛型載具影像數值地形模型建置及精度評估,碩士論文,國立臺北科技大學土木與防災研究所,台北,2013。
[15] 張翊晨,應用遙感探測與野外現地測量初探花蓮光復地區地表變形,碩士論文,國立臺北科技大學土木與防災研究所,台北,2013。

延伸閱讀