(34.239.176.198) 您好!臺灣時間:2021/04/23 20:00
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果

詳目顯示:::

我願授權國圖
: 
twitterline
研究生:葛士豪
研究生(外文):Shyh-Haur Ger
論文名稱:植基於主從式網路架構的行動裝置三次元繪圖系統
論文名稱(外文):A Client-Server Framework for 3D Graphics on Mobile Devices
指導教授:張鈞法
指導教授(外文):Chun-Fa Chang
學位類別:碩士
校院名稱:國立清華大學
系所名稱:資訊工程學系
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:2003
畢業學年度:91
語文別:英文
中文關鍵詞:以影像為基礎的貼圖行動裝置影像重建定點數運算主從式網路架構
外文關鍵詞:3d image warpingsplattingfixed-point arithmeticmobile devicesclient-serverGAPI
相關次數:
  • 被引用被引用:0
  • 點閱點閱:61
  • 評分評分:系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
由於行動裝置的逐漸普及,使用者對於行動裝置的能力的要求也更形廣泛.其中也包括了三次元繪圖的能力.有些人會使用如miniGL[4]或PocketGL[6]這類的函式庫來開發行動裝置上的三次元繪圖應用程式,不過因為這類的函式庫是將傳統的繪圖管線直接移植到行動裝置上,又因為行動裝置的運算能力不強,因此貼圖速度也不高.此外,當場景複雜時所需用以存放模型資料的記憶體需求也將是一大問題.
在這篇論文中,我們將提出一個植基於主從式網路架構的行動裝置三次元繪圖系統.在這套系統中,在pc上的伺服器將負責作傳統的高品質貼圖,並將貼圖結果傳給客戶端.客戶端收到後,便使用一種以影像為基礎的貼圖法來作後續的貼圖動作.
然而,在客戶端的貼圖方法也將遇到影像品質的問題.我們將使用一種模擬影像中的像素如何被周圍像素影響的演算法來修補這類的瑕疵.此外,由於行動裝置並未有硬體浮點運算的功能,我們也將使用定點數的算數來取代浮點數的算數.
我們的實驗平台是架構在Microsoft WinCE based PocketPC上,在輸出影像尺寸為240*180的情況下,貼圖的速度大約是每秒4-5張.

3D Graphics on mobile devices has been an arising requirement in recent years. Some application programs are developed with graphics library like miniGL[4] or PocketGL[6], which are based on traditional graphics pipeline. Due to insufficient computation power of mobile devices, it is difficult to run traditional graphics pipeline as smooth as on desktop PC. The memory capacity presents another problem when holding large scale models.
In this paper, we propose a client-server based rendering mechanism to replace the traditional graphics pipeline. Within this mechanism, the scene (model) is rendered on desktop PC (server) by OpenGL[5], then the rendering result is sent to mobile devices (client). The post-rendering on mobile devices is constructed by an image based rendering(IBR) method, which is called 3D Image Warping[2][3].
Artifacts may occur when using 3D image warping. To solve this, we also implement a simple splatting[7][9] method which consumes a little more extra computation power but provides satisfactory image quality. Besides, to improve frame rate, floating-point numbers computations are replaced by fixed-point computations.
The mobile devices used in our system are Microsoft WinCE based PocketPC, with 206 Mhz StrongArm processors. The frame rate is about 4.5 when splatting is off and 4 when on for output image size of 240 by 180.

Index
List of figures……………………………………………………………3
Abstract………………………………………………………………….4
1 Introduction……………………………………………………….....5
2.Problem Analysis…………………………………………………….7
Platform Characteristic…………………………………………………7
Solution……………………………………………………………………7
3.Post-Rendering on Client……………………………………………9
3Dwarping……………………9
The 3D warping model………………………………………………….9
Eliminate artifacts…………………………………………………11
Visibility problem and scan direction determination…… 13
Fixed-point arithmetic…………………………………………….16
4.Pre-Rendering on Server………………………………………..18
Z-value disparity correspondence……………………………...18
5. System Architecture…………………………………………...21
5.1 System architecture………………………………………21
5.2 Network communication issue………………………..21
5.3Drawing with Game API……………………………………………22
6. Experimental Results……………………………………………….24
7. Conclusion and Future Work………………………………………26
8. Acknowledgement…………………………………………………...27
9. References……………………………………………………………27
Appendix A……………………………………………………………..29

1. The Game API website: http://www.pocketpcdn.com/sections/gapi.html
2. Leonard McMillan and Gary Bishop. “Plenoptic Modeling: An image-based rendering system”.In SIGGRAPH 95 Conference Proceedings, pages 39—46, August 1995.
3. Leonard McMillan. An Image-Based Approach to Three-Dimensional Computer Graphics.Ph.D. Dissertation. Technical Report 97-013, University of North Carolina at Chapel Hill, Department of Computer Science, 1997.
4. MiniGL by Digital Sandbox, Inc. The miniGL website: http://www.dsbox.com/minigl.html
5. OpenGL website: http://www.opengl.org
6. PocketGL website: http://www.sundialsoft.freeserve.co.uk/pgl.htm
7. Szymon Rusinkiewicz and Marc Levoy. “QSplat: A Multiresoluton Point Rendering System for Large Meshes”. In SIGGRAPH 2000 Conference Proceedings, pages 343—352, July 2000.
8. Jonathan Shade, Steven Gortler, Li-wei He and Richard Szeliski. “Layered Depth images”. In SIGGRAPH 98 Conference Proceedings, pages 231—242, July 1998.
9. Lee Westover. SPLATTING: A Parallel, Feed-Forward Volume Rendering Algorithm. Ph.D. Dissertation. Technical Report 91-029, University of North Carolina at Chapel Hill.

QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
系統版面圖檔 系統版面圖檔