進階搜尋


下載電子全文  
系統識別號 U0026-0812200915331200
論文名稱(中文) 應用移動式測圖系統影像序列萃取三維道路邊線及分隔線
論文名稱(英文) Extraction of 3D Road Boundaries and Dividing Lines Using MMS Image Sequences
校院名稱 成功大學
系所名稱(中) 測量及空間資訊學系碩博士班
系所名稱(英) Department of Geomatics
學年度 98
學期 1
出版年 98
研究生(中文) 李荏毓
研究生(英文) Ren-Yu LI
學號 p6696402
學位類別 碩士
語文別 英文
論文頁數 68頁
口試委員 指導教授-曾義星
口試委員-陳良健
口試委員-史天元
中文關鍵字 移動式測圖系統  多時序影像  道路邊線  中心線 
英文關鍵字 Road Boundaries  Dividing Lines  MMS Image Sequences  Mobile Mapping 
學科別分類
中文摘要 路網資料對於許多層面例如城市規劃、交通規劃以及導航的應用上均有相當大的重要性。道路線型資料是道路表示的基本要素,擷取二維及三維到路線型資料便為維護道路資料的重要工作。常見傳統道路資料擷取方法為航空攝影測量與人力調查,航空攝影測量成本過高而人力測量所及範圍有限,均無法達到最佳的效率,因應資料獲取無效率的問題,近年發展了移動式測圖系統。移動式測圖系統裝載率定過的數位相機,能及時拍攝多時序影像且具有方位資訊的立體像對。為此,本研究以移動式測圖系統作為研究平台,提出半自動化萃取三維道路線型的方法。
本研究結合邊線偵測與模板匹配來萃取影像中坐落於道路邊線或中心線上的像點。運用核線幾何概念加上針對不同像對間匹配的尺度調整,各重疊像對中的共軛點位置即可被定出。利用多個共軛點位置做為多於觀測,精確的三維物點坐標即能被獲得。為自動萃取道路線型,在自動化擷取上利用B-spline於二維空間以及三維空間線型套合(B-spline curve fitting)後所提供的資訊提供模板匹配需要的角度資訊,以及車體前進方向的粗估,除此之外,本研究也將B-spline的特性納入研究中作為濾除大錯誤的機制。為證實本演算法可行度,本實驗測試了36份移動式測圖系統的多時序像對,結果顯示本演算法能成功自動擷取道路邊線與中心線。額外,本研究也針對不同型態的台灣道路影像做測試,以提供分析探討依據。
英文摘要 Road network data is essential for many applications, such as urban planning, transportation management and car navigation. Detail 2-D and 3-D road line data are required to provide and maintain the up-to-date data service. Conventional approaches for data acquisition neither time effective nor cost effective, an alternative data collection system was developed to overcome the disadvantages of above methods. A mobile mapping system (MMS) equipped with digital cameras captures orientation-known image sequences along a road path. The fast data collection speed and the mobility along with the background idea of photogrammetry, an MMS can largely increase the efficiency for data maintenance. In this study, we proposed a semi-automatic method to extract 3-D road line from MMS image sequences.

Edge extraction combining with template matching is performed to extract edges located on road boundaries and center lines from MMS images. 3-D object coordinate of road image points are obtained by finding conjugate points in overlapped images through the use of epipolar geometry and image matching as well as scale adjustments among image sequences. Our algorithm performs automatic extraction by determining rotation angle for template matching and approximate extension of road line section in next image sequence by implemented B-spline curve fitting. We also use the characteristic of B-spline in detecting blunders. Our algorithm was tested with 36-pair MMS image sequences and results show that our algorithm can successfully extract 3-D road lines. Different road cases in Taiwan are also tested to evaluate the performances of road line extractions with different road images.
論文目次 CHINESE ABSTRACT I
ENGLISH ABSTRACT III
ACKNOWLEDGEMENT IV
LIST OF TABLES VII
LIST OF FIGURES VIII

CHAPTER1 INTRODUCTION
1-1 BACKGROUND AND MOTIVATION 1
1-2 PREVIOUS STUDIES 2
1-3 RESEARCH APPROACH 3
1-3.1 MMS IMAGES 3
1-3.2 RESEARCH APPROACH 5
1-3.3 ROAD LINES EXTRACTION 7
1-3.4 DATA POINTS DETERMINATION 8
1-3.5 NORMALIZED CROSS-CORRELATION AND TEMPLATE MATCHING 11
1-4 THESIS STRUCTURE 11

CHAPTER2 3-D COORDINATE DETERMINATION
2-1 EPIPOLAR GEOMETRY 12
2-2 MATCHING STRATEGY AND 3-D COORDINATE DETERMINATION 13
2-3 SCALE VARIATION AND RELIABILITY IMPROVEMENT 13
2-3.1 SCALE VARIATION 14
2-3.2 RELIABILITY IMPROVEMENT19
2-4 THE DETERMINATION OF 3-D DATA POINT COORDINATES 23
2-4.1 TEMPLATE SHAPE 23
2-4.2 COORDINATE DETERMINATION 25

CHAPTER3 B-SPLINE CURVE FITTING
3-1 Quadratic Uniform B-SPLINE 27
3-2 LEAST SQUARE B-SPLINE CURVE FITTING 28
3-3 TRANSITION FROM 2-D TO 3-D 29
3-3.1 TEMPLATE MATCHING 29
3-3.2 BLUNDER DETECTION 30
3-3.3 EXTENSION OF ROAD LINE SECTION 31
3-4 TERMINATION OF RECURSIVE PROCESS 32

CHAPTER4. RESULTS AND ANALYSES
4-1 TEST DATA 33
4-2 EXTRACTION OF 2-D ROAD LINES 33
4-2.1 EDGE DETECTION 34
4-2.2 EXTRACTION OF 2-D ROAD LINES 39
4-3 EXTRACTION OF 3-D ROAD LINES 43
4-3.1 CONJUGATE POINT DETERMINATION 43
4-3.2 B-SPLINE CURVE FITTING 46
4-4 BLUNDER DETECTION 54

CHAPTER5. DISCUSSION AND SUGGESTIONS 62

CHAPTER6. CONCLUSIONS 66

REFERENCES 67
參考文獻 Gonzalez. R.C.& Woods. R.E., Digital Image Processing, pp. 80-84, 519-550, 2002.
Haihonng Li, Semi-automatic Road Extraction from Satellite Images, PhD dissertation, 1997.
Naser El-Sheimy, The development of VISAT – A Mobile Survey System For GIS Application, PhD dissertation, 1996.
Rogers. D. F. An introduction to NURBS with Historical Perspective, 2001

Asif M. Arshad M.R., Zia M.Y.I. & Yahya A., “An implementation of active contour and Kalman filter for road tracking” Internatioal Journal of Applied Mathematics, Vol. 37(2), 2007.
Canny. J. “A Computational Approach to Edge Detection” IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, Vol. 8, pp. 679-698, 1986.
deBoor. C. “On Calculating B-spline” Journal of Approximation Theory, Vol. 6, pp. 50-62, 1972.
Gontran. H., Skaloud. J. & Gilliéron. P.Y., “A Mobile Mapping System for Road Data Capture via a Single Camera” Infoscience, 2003.
He G. P. & Novak K., “On line data capture by image matching on the mobile mapping system” SPIE, Vol. 1820, pp. 50-56, 1992.
Hu. X., Zhang. Z. & Tao. C. V., “A Robust Method for Semi-Automatic Extraction of Road Centerlines Using a Piecewise Parabolic Model and Least Square Template Matching” PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING, Vol.70 (12), pp. 1393-1398, 2004.
Lin. X., Zhang. J., Liu. Z. & Shen. J., “Integration method of profile matching and template matching for road extraction from high resolution remotely sensed imagery” IEEE, 2008.
Shao J. L. & Clive F., “Three-image matching for 3-D linear object tracking” Geo-spatial Information Science, Vol. 3(2), pp. 13-18, 2000.
Shiode N., “3D urban models: Recent developments in the digital modeling of urban environments in three-dimensions” GeoJournal, Vol. 52, pp. 263–269, 2001.
Sawano H. & Okada M., “A road extraction method by an active contour model with inertia and difference features”, IEICE TRANS. INF & SYS, Vol. E89-D (7), pp. 2257-2267,2006.
Tao C.V., Chapman M.A. & Chaplin B.A., “Automated processing of mobile mapping image sequences” Photogrammetry & Remote Sensing, Vol. 55, pp. 300-346, 2001.
Wang Y., Teoh E. K. & Shen D., “Lane detection and tracking using B-Snake” Image and Vision Computing, Vol. 22, pp. 269-280, 2004.
Wang W., Pottmann H. & Liu Y., “Fitting B-spline curves to point clouds by curvature-based squared distance minimization” ACM Transactions on Graphics, Vol. 25(2), pp. 214–238, 2006.
論文全文使用權限
  • 同意授權校內瀏覽/列印電子全文服務,於2009-12-11起公開。
  • 同意授權校外瀏覽/列印電子全文服務,於2009-12-11起公開。


  • 如您有疑問,請聯絡圖書館
    聯絡電話:(06)2757575#65773
    聯絡E-mail:etds@email.ncku.edu.tw