政大機構典藏-National Chengchi University Institutional Repository(NCCUR):Item 140.119/55478
English  |  正體中文  |  简体中文  |  Post-Print筆數 : 27 |  Items with full text/Total items : 113656/144643 (79%)
Visitors : 51747650      Online Users : 591
RC Version 6.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
    •  Adv. Search
    Home ‧ Login ‧ Upload ‧ Help ‧ About ‧ Administer Goto mobile version


    Please use this identifier to cite or link to this item: https://nccur.lib.nccu.edu.tw/handle/140.119/55478


    Title: 應用於地理資訊系統之使用多視角影像重建三維建物模型
    Other Titles: 3d Building Model Reconstruction Using Multi-View Images for GIS Applications
    Authors: 何瑁鎧;唐政元
    Contributors: 國立政治大學資訊科學系
    行政院國家科學委員會
    Keywords: Bundle Adjustment;Epipolar Transfer;三維建物模型重建;多視角影像;地理資訊系統
    Bundle Adjustment;Epipolar Transfer;3D Building Model Reconstruction;Multi-View Images;GIS
    Date: 2010
    Issue Date: 2012-11-12 11:05:18 (UTC+8)
    Abstract: 近年來由於資訊技術迅速成長，網路普及與頻寬提升，資訊科技的應用進入前所未有的新領域，直接或間接的推動傳統的地理資訊系統(GIS: Geographic Information System) 由二維往三維發展，而數位地球(digital earth)與數碼城市(cyber city)等相關概念與研究應運而生，如何有效、精確且迅速的建構數碼城市等研究日趨重要。建構數碼城市，除了要建立環境中所有建築物的三維立體模型外，同時還需考慮與地理資訊系統的結合，才能在日後方便的對城市作分析、規劃與導覽等研究。在此前提下，我們擬定了本計畫，希望能提供一個有效、精確且迅速的方式來建立三維建物模型，並與地理資訊系統結合以呈現之。計畫內容主要就三維之建物模型建構進行研究，我們期望利用低成本的數位相機，拍攝與蒐集多視角建物影像，進行建物模型重建。主要的研究方法大抵可分為下列步驟：首先我們利用Harris filter 與Difference of Gaussian 等方法自動偵測各影像中稀疏對應點；其次我們利用稀疏對應點結合三維幾何關係進行擴展，產生大量密集對應點；再者，我們將利用極線轉換關係濾除與改善錯誤的對應後，再利用bundle adjustment 將投影誤差最小化，並經過多次遞回處理，進行相機重新校正與三維點雲最佳化；最後利用前述結果的點雲資料重建建物模型，其中我們採用三維補綴面投影之觀念，以3D 幾何和關係來改善對應點的相似度計算。同時自動從多視角影像中，合成出較佳的敷貼材質，貼附於模型表面上，使建物模型更接近於真實建築物。我們預期在校園內拍攝大量多視角影像進行實驗，利用上述方法重建校園內之三維建物模型，並將所建立之模型與地理資訊系統結合，實作出三維校園導覽系統。我們也預計將所建立之模型與虛擬實境導覽結合，進一步以虛擬實境技術呈現校園導覽。目前在實驗實作上已有初步的成果，期望藉由此研究結果，未來能與各研究領域進行整合，以期對地理資訊系統、都市計畫、防災妨害與虛擬實境導覽等方面有皆所貢獻。
    The rapid advances in information technologies and network popularity as well as its bandwidth increasing lead the applications of information technologies into a new era. These advancements led the researches and applications of the traditional Geographic Information System (GIS) from two dimensional into three dimensional. The concepts and researches of Digital Earth as well as Cyber City come with the tide of fashion. Hence, the researches on efficiently and accurately construct a Cyber City gradually becomes one of the major issues. In constructing a Cyber City, one has to consider not only the reconstruction of all the building models in the environments but also the integration with geographic information system in order to facilitate future planning, analysis, and navigation in the Cyber City. Under these concepts, we are writing this proposal to provide an efficient, accurate, and rapid mechanism in three-dimensional building model reconstruction for GIS. In this project, we focus on the reconstruction of three-dimensional building models. Our goal is to produce/collect a large set of multi-view images about the target buildings using low-cost digital cameras. Using these images as inputs, we reconstruct the 3D models of these buildings through the following steps. In the first step, we detect automatically a sparse set of corresponding points using the Harris filter and Difference of Gaussian filter. In the second step, we can use the sparse corresponding points and properties of 3D geometries to generate a quasi-dense set of corresponding points. And then we can apply the epipolar transfer to remove incorrect corresponding points. After that, we can reduce the projection error by applying standard bundle adjustment algorithm repeatedly and get the improved camera parameters and 3D point cloud. Finally, the three-dimensional build models can be reconstructed from the good corresponding points. Refer to the similarity measurement, the 3D patch projection is used to improve the results of correspondence. Furthermore, we can synthesize the texture from the multi-view images and apply it to the building models to make the model more realistic. We plan to take a large set of multi-view images of our campus as the experimental data and apply our mechanism to reconstruct the 3D building models of our campus. We will integrate our results with geographic information system to generate a realistic campus model. One may hope to conduct a virtual reality system using our model for campus navigation. The current stage, we have some preliminary results and hopefully can integrate these results with others. In the future, we plan to extend our idea to co-operate with other researchers in different area in order to promote this idea for city planning, disaster prevention, as well as others.
    Relation: 基礎研究
    學術補助
    研究期間:9908~ 10007
    研究經費:604仟元
    Data Type: report
    Appears in Collections:[Department of Computer Science ] NSC Projects

    Files in This Item:

    File Description SizeFormat
    99-2221-E-004-011.pdf849KbAdobe PDF2861View/Open


    All items in 政大典藏 are protected by copyright, with all rights reserved.


    社群 sharing

    著作權政策宣告 Copyright Announcement
    1.本網站之數位內容為國立政治大學所收錄之機構典藏，無償提供學術研究與公眾教育等公益性使用，惟仍請適度，合理使用本網站之內容，以尊重著作權人之權益。商業上之利用，則請先取得著作權人之授權。
    The digital content of this website is part of National Chengchi University Institutional Repository. It provides free access to academic research and public education for non-commercial use. Please utilize it in a proper and reasonable manner and respect the rights of copyright owners. For commercial use, please obtain authorization from the copyright owner in advance.

    2.本網站之製作，已盡力防止侵害著作權人之權益，如仍發現本網站之數位內容有侵害著作權人權益情事者，請權利人通知本網站維護人員(nccur@nccu.edu.tw)，維護人員將立即採取移除該數位著作等補救措施。
    NCCU Institutional Repository is made to protect the interests of copyright owners. If you believe that any material on the website infringes copyright, please contact our staff(nccur@nccu.edu.tw). We will remove the work from the repository and investigate your claim.
    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback