由於自走車的發展日新月異，人類對自走車的應用也越來越廣泛，從自動化的工廠搬運到日常生活的屋內打掃，都離不開自走車的應用範圍。為了使自走車能夠在未知的環境中自由行走，必須建立一個可靠的智慧型導航系統，使自走車對周圍環境有正確的認知，以實現在操作環境中避障的判斷能力。
在本論文中，我們介紹同軸幾何來說明雙眼立體視覺的幾何關係，並且解釋一般的雙眼視覺之基本特徵點尋找與對應點比對的問題。在障礙物特徵點或外型的偵測方面，則是採用Planar Homography所延伸出來的“平面衍生視差”，利用不屬於某特定平面上的物體便會產生視差的特性，將障礙物與地面交界處的輪廓擷取出來，當作計算障礙物與自走車之間距離的對應點，使結果能夠表示出障礙物的外型與距離。最後將攝影機作不同角度的拍攝，並且將障礙物與自走車的相關位置合併起來，形成自走車周圍的環境資訊，以利自走車作初始位置的路徑判斷。

Due to the development of Autonomous Mobile Robot (AMR), the AMR is applied widely. The applications of the AMR include the works from transport in the automatic factory to household cleaning in our daily life. In order to move arbitrarily in the unknown environment, a reliable intelligent navigation system for the AMR should be established. This system will make the AMR to get correct information on implementing the capability of avoidance obstacle in the control environment.
In this thesis, Epipolar Geometry is introduced to explain the geometric relation of binocular stereo vision. The works of extracting the feature points and matching corresponding points in binocular vision are explained. In the aspect of features and shape of the obstacle, the concept of plane induced parallax from planar homography is applied to collect the border between the obstacle and ground. These borders will be the feature points of the obstacle for calculating the distance from the AMR to the obstacle. Lastly, the images shot at different angles would be use to integrate the position of the obstacle. These results display the environment information around the AMR. It will be useful for the AMR to determine the path.

[1] M.Bertozzi and A Broggi,“GOLD: A Parallel Real-Time Stereo Vision System for Generic Obstacle and Lane Detection.” IEEE Transaction on Image Processing, Vol 7, No1, January 1998.

[2] Q.-T. Luong, J. Weber, D. Koller and J. Malik, “An integrated stereo-based approach to automatic vehicle guidance”, In 5th international Conference on Computer Vision, June 1995.

[3] H. Li and Simon X. Yang, “Ultrasonic Sensor based Fuzzy Obstacle Avoidance Behaviors”, IEEE SMC MP2L2, 2002.

[4] Cheng-Chih Lin and R. Lal Tummala, “Adaptive Sensor Integration for Mobile Robot Navigation”, IEEE International Conference on Multisensor Fusion and Integration for Intelligent System, Oct 1994.

[5] E. Elkonyaly, F Areed, Y. Enab and F. Zada, “Range sensory-based navigation in unknown terrains”, Proc. SPIE, Vol.2591, 1995.

[6] http://millemiglia.ce.unipr.it/ARGO/english/techdet/vd.html, Vehicle Detection in ARGO System.

[7] R. Hartley and A. Zisserman, Multiple View Geometry in Computer Vision, Cambridge University Press, Cambridge, 2000.

[8] R. Klette, K. Schluns and A. Koschan, Computer Vision : Three-Dimensional Data from Image, Springer-Verlag, Singapore, 1998.

[9] E. Trucco and A. Verri, Introductory Techniques for 3-D Computer Vision, Prentice-Hall, New Jersey, 1998.

[10] J. Shi and C. Tomasi, “Good Features to Track”, In IEEE Computer Society Conference on Computer Vision and Pattern Recognition, June 1994.

[11] E. Vincent and R. Laganiere, “Detecting Planar Homographies in an Image Pair”, Proc. 2nd International Symposium on Image and Signal Processing and Analysis, pp. 182-187, (Pula, Croatia,) June 2001.

[12] R. Labayrade, D. Aubert and J.-P. Tarel, “Real Time Obstacle Detection in Stereovision on Non Flat Road Geometry Through “V-disparity” Representation”, IEEE Intelligent Vehicle Symposium, June 2002.

[13] L. Fletcher, N. Apostoloff, J. Chen and A. Zelinsky, “Computer Vision for Vehicle Monitoring and Control”, Australian Conference on Robotics and Automation, Sydney, Australia, 2001 (ACRA 2001).

[14] Z. Zhang, “A Flexible New Technique for Camera Calibration”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol 22, No 11, November 2000.

[15] R. Gonzalez and R. Woods, Digital Image Processing, Prentice-Hall, New Jersey, 2002.

[16] 陳饒中, 自走車系統之障礙物檢測與位置資料之建立, 國立成功大學, 碩士論文, 民國90年。

[17] 吳至仁, 即時障礙物偵測/定位及標誌辨識, 國立成功大學, 碩士論文, 民國91年。

[18] 呂其展, 利用影像序列建構與顯示三維地形模型之研究, 國立成功大學, 碩士論文, 民國91年。

[19] 孫正泰, 以電腦視覺為基礎之道路上障礙物偵測, 國立臺灣大學, 碩士論文, 民國90年。

[20] 劉睿智, 利用多隻相機之電腦視覺與自動車航行技術作建築物走廊之安全巡邏, 國立交通大學, 碩士論文, 民國89年。

[21] 黃聿彣, 三維空間多攝影機視覺監控系統, 國立交通大學, 碩士論文, 民國91年。