以自動化方式進行立體製圖時，常遇到如遮蔽、陰影或特徵不明顯等資訊不足的情況，此時須將人為判斷與推論的方式與過程程式化。為了減少此種資訊不足的情況，常透過增加幾何、光譜的資訊或加入人為的判斷，來輔助人工建築物特徵萃取或三維模型重建的工作。
考量幾何資訊之增加，本研究嘗試利用雙向重疊均約為60%的四重疊影像，進行房屋三維模型的重建工作。主要有二個處理程序，（1）最小二乘法影像匹配：產生數值表面模型（Digital Surface Model）及（2）一序列的處理程序，包括了最鄰近勻化法、多重解析力邊緣偵測、局部最大值邊緣細化、CLR匹配法。利用數值表面模型得到房屋候選區塊，找出每一個房屋候選區塊內的房屋邊界候選點，並進行Hough直線偵測、直線分組與屋角點三維座標之計算。以上處理之各雙像立體對亦完成整合。
由實驗結果得知，多重疊航測影像可萃取出較多房屋邊界的資訊，並經由立體匹配得到房屋邊界之三維資訊，進而重建房屋三維模型。另外，在某些條件下，斜頂房屋物空間之缺漏邊界，亦能透過物空間與像空間的轉換，到像空間去搜尋是否有符合條件的直線段，輔助人工建築物三維模型重建工作。

In an automated procedure for stereoscopic mapping, reasoning is always needed when insufficient information due to occlusions, shadows, or indistinctness of image features is encountered. In order to alleviate the influence of the information insufficiency, it is common to include more geometric or spectral data source to assist the reconstruction for 3-D man-made buildings.
Considering the improvement of geometric information, this investigation uses aerial quadruplets with 60% overlap in two perpendicular directions to perform the reconstruction for man-made buildings. Two major components include (1) generation of digital surface model (DSM) by templet matching, and (2) feature extraction and processing such as image smoothing, edge detection, edge thinning, and CLR matching. Based on the blocks extraction from DSM, each potential building is extracted followed by edge detection, Hough transform to detect straight lines, line grouping, and 3-D positioning for building corners. All the stereopairs composed by 4 images are integrated in this study.
Experimental results indicate that, as compared to stereopairs quadruplets could be extracted more edge information for better 3-D building reconstruction. In addition, it is also possible to complete the gable roof when only partial edge information is available by using dual searching between image space and object space.

目 錄
中文摘要 …………………………………………..…………I
英文摘要 ………………………………………..…………...II
誌謝 ………………….….………………….………………...III
目錄 ………………………………………………….…...…...Ⅳ
圖目錄 ……………………………………………………...VII
表目錄 …………………………………………………….…XI
第一章 緒論
1-1 研究動機及目的 …………………………………………1
1-2 文獻回顧 …………………………………………………3
1-3 研究構思與方法 …………………………………………6
1-4 論文架構 …………………………………………………9
第二章 特徵萃取與影像匹配
2-1 特徵之意義 ……………………………………………..12
2-2 影像之前處理 …………………………………………..12
2-3 邊緣線偵測 ……………………………………………..14
2-4 邊緣線之細化 ……….………………………………….17
2-5 CLR匹配法 …….….………………………………….18
2-6 數值表面模型 …….….………………………………..22
第三章 人工建築物三維模型之重建
3-1 人工建築物候選區塊 …….…………………………….25
3-2 人工建築物候選直線邊界 .…………………………….25
3-3 重建人工建築物三維模型 …….……………………….30
3-3-1 直線分組 …….…………………………………..31
3-3-2屋角點物空間三維座標之計算 …..……………..32
3-4 斜頂房屋缺漏邊界之處理 …….……………………….38
第四章 實驗與成果
4-1 實驗資料簡介 …….…………………………………….42
4-2 測試例一 …….………………………………………….49
4-2-1 人工量測得到之屋角點三維座標 ………….…..49
4-2-2 邊緣線萃取成果 ………………….……….…….52
4-2-2-1 SNN-Mean Filter所得之成果 ………….52
4-2-2-2 多重解析力邊緣偵測成果 ………….....52
4-2-2-3 局部最大值追蹤邊緣細化法成果 …….53
4-2-3 以層次式影像匹配得到的成果 ………………...60
4-2-4 房屋候選區塊成果 ……………….………...…...61
4-2-5 房屋候選邊界點成果 …………….……………..62
4-2-6 人工建築物候選直線邊界成果 ………………...64
4-2-7屋角點物空間三維座標計算成果 ………….…...65
4-2-8 處理斜頂房屋缺漏邊界之成果 ……………..….70
4-3 測試例二 ………………………..………..………….….79
4-3-1 邊緣線萃取成果 ……………………………..….79
4-3-2 DSM與房屋候選區塊成果 ……….………..….80
4-3-3房屋候選邊界點成果 ……………….………..….81
4-3-4 人工建築物三維模型重建成果 ……………..….83
第五章 結論與建議 ……………………………………....87
參考文獻 …………………………………………………....89

參 考 文 獻
李松釧，「SPOT衛星立體像對的匹配與精度之分析」，國立中央大學大氣物理研究所碩士論文，中壢，1988。
吳碧玉，「以二維CAD數值圖輔助屋頂模型之重建」，國立成功大學測量工程研究所碩士論文，台南，1998。
徐百輝，「應用Hough轉換於框標及共軛影像量測之研究」，國立成功大學測量工程研究所碩士論文，台南，1995。
徐偉城，「空照彩色立體像對中人工建築物萃取之研究」，國立中央大學土木工程研究所碩士論文，中壢，1999。
許才仁，「應用局部過濾及邊緣線對應改善數化空照立體對之區域匹配」，國立中央大學土木工程研究所碩士論文，中壢，1996。
陳慶鵬,「數位空照立體對邊緣萃取與匹配之研究」,國立中央大學土木工程研究所碩士論文，中壢，1996。
黃智仁，「影像特徵萃取與立體匹配之結合於人工建物偵測之研究」， 國立中央大學土木工程研究所碩士論文，中壢，1998。
蔡玉慧，「以色彩資訊協助影像特徵之萃取」，國立成功大學測量工程研究所碩士論文，台南，1998。
饒見有、陳良健，「紋理分析於SPOT立體像對影像匹配動態視窗之應用」，第十五屆測量學術及應用研討會，第717-725頁，1996。
饒見有、陳良健，「局部最大值追蹤法於邊緣影像細化及向量化之研究」，八十六年電子計算機於土木水利工程應用論文研討會論文集，第471-476頁，1997。
饒見有、陳良健，「利用Hough轉換於直線之主要結構分析與偵測」，第十七屆測量學術及應用研討會，第983-992頁，1998。
Baillard,C.、Dissard, O.、Jamet,O.and Maitre, H. , “Extraction and textural characterization of above-ground areas from aerial stereo pairs : a quality assessment”, ISPRS Journal of Photogrammetry & Remote Sensing,Vol.53,pp.130-141,1998.
Baitsavias,E.P. and Mason,S.O. ,“Automated Shack Reconstruction Using Intergration of Cues in Object Space”,IAPRS,3D Reconstruction and Modeling of Topographic Objects，Vol.32，Part3-4W2,pp.96-105,1997.
Bignone,F. ,“Segment Stereo-Matching and Coplanar Grouping” , Technical Report BIWI-TR-165,Institute of Communications Technology,Image Science Lab,ETH,Zurich,Switzerland,pp.1-30,1995.
Canny,J., “A Computational Approach to Edge Detection”, IEEE Transactions of PAMI ,Vol. PAMI-8,No.6,pp.679-698,1986.
Chen,L.C. and Lee,L.H. ,“Rigorous Generation of Digital Orthophoto from SPOT Images”,Photogrammetric Engineering and Remote Sensing,Vol.59,No.5,pp.655-661,1993.
Deok, J. P., Nam,K. M. and Park, R. H., “Multi-resolution edge detection techniques”,Pattern Recognition,Vol.28,No.2,pp.211-229,1995.
,W.,“Quality Assessment of Object Location and Point Transfer Using Digital Image Correlation”,ISPRS Congress,Commission III,Rio,de Janeiro,pp.197-219,1984.
Gabet, L., Giraudon, G., and Renouard, L., “Automatic Generation of High Resolution Urban Zone Digital Elevation Models”, ISPRS Journal of Photogrammetry and Remote Sensing,Vol. 52,No. 4, pp.33-47,1997.
Halla,N. and Brenner,C. ,“Interpretation of Urban Surface Models Using 2D Building Information”,Automatic Extraction of Man-Made Objects from Aerial and Space Images (II),Birkhauser Verlag,Berlin,pp.213-222,1997.
Halla,N. and Brenner,C. ,“Interpretation of Urban Surface Models Using 2D Building Information”, Computer Vision and Image Understanding,Vol.72,No.2,pp.204-214,1998.
Harwood,D., Subbarao,M., Hakalahti, H. and Davis, L. S., “A new class of edge-preserving smoothing filters”,Pattern Recognition Letters,Vol.6,pp.155-162 ,1987.
Henricsson,O.,Bignone,F.,Willuhn,W.,Aef,F.,Kuebler,O.,Baltsavisa,E.,Mason,S, and Greun,A.,“PROJECT AMBOE: Strategies,Current Status and Future Work”, ISPRS Congress, Commission III , Vol.31,Part B3,pp.321-330,1996.
Henricsson,O. and Baltsavias,E.,“3-D Building Reconstruction with ARUBA:A Qualitative and Quantitative Evaluation”,Automatic Extraction of Man-Made Objects from Aerial and Space Images（II）,Birkhauser Verlag,Berlin,pp.65-76,1997.
Henricsson,O. ,“The Role of Color Attributes and Similarity Grouping in 3-D Building Reconstruction”,Computer Vision and Image Understanding,Vol.72,No.2,pp.163-184,1998.
Hough,P.V.C. ,“Methods and Means for Recognising Complex Patterns”, U.S. patent NO.3069654,1962.
Kim,T. and Muller,J.-P ,“A Technique for 3D Building Reconstruction”, Photogrammetric Engineering and Remote Sensing,Vol.64, No.9,pp.923-930,1998.
Leavers,V.F. ,“Shape Detection in Computer Vision Using the Hough Transform”, Springer-Verlag London Limited,1992.
Li,J.,Nevatia,R. and Nornoha,S. , “User Assisted Modeling of Buildings”, Internal Report,http://iris.usc.edu,1999.
Risse,Y. ,“Hough Transform for Line Recognition: Complexity of Evidence Accumulation and Cluster Detection”,Computer Vision Graphics and Image Processing,Vol.46,pp.327-345, 1989.
Rosenholm,D. ,“Multi-Point Matching Using the Least Squares Technique for Evaluation of Three Dimensional Models”, Photogrammetric Engineering and Remote Sensing,Vol.53, No.6,pp.621-626,1987.
Sahar,L. and Krupnik,A.,“Semiautomatic Extraction of Building Outlines from Large-Scale Aerial Images”,Photogrammetric Engineering and Remote Sensing,Vol.65,No.4,pp.459-465,1999.
Spreeuwers,L. , Schutte,K. and Houkes,Z.,“A Model Driven Approach to Extract Buildings from Multi-view Aerial Imagery”,Automatic Extraction of Man-Made Objects from Aerial and Space Images (II),Birkhauser Verlag,Berlin,pp.109-118,1997.
Tudor,G.S. and Sugarbaker,L.J. ,“GIS Orthographic Digitizing of Aerial Photographs by Terrain Modeling”,Photogrammetric Engineering and Remote Sensing ,Vol.59, pp.499-503,1993.