在電腦視覺裡，三維空間上的場景的重建需要藉由影像對應求得其三維座標。但在一些影像特徵不明顯的區域，比如單一色調的景象。或是影像重覆性很高的地方，如有重覆圖案的物體。很容易找到錯的影像對應。因此無法提供正確的三維空間資訊。在我們觀察中，這些場景通常具有平面的特性。
考慮到上述的情況，本篇論文提出一個在多平面場景環境下的影像對應方法。首先求得影像中各個平面的單應性矩陣，再將影像中各個平面的位置切割出來。將二張影像經一投影轉換矯正後，使得對應點的搜尋由二維變成一維。藉由知道單應性矩陣和平面位置，我們可以求得包含在這個平面的每個像素所對應到另一張影像中的像素位置，其結果將以視差圖表示。

In computer vision, 3D reconstruction needs 3D positions which are obtained by stereo matching. The mismatching would happen in the region of no features such like the scene with smooth color distribution, or region of repeating patterns. Therefore 3D information might be wrong. According to our observation, most of these scenes are planar.
Considering the above points, in this thesis, a method of stereo matching of planar scenes is proposed. First, the homography of each is obtained and then we use homograpy to segment each plane. After that, two images have to be rectified so that the corresponding search would change from 2D to 1D. Because the homography and region of plane has been obtained, the approximate position can be known by transforming point in view one to view two. The result would be shown in a dense disparity map.

[1] M. Han and T. Kanade: Creating 3D Models with Uncalibrated. Cameras. Proc. Workshop Applications of Computer Vision, pp. 178-. 185, 2000

[2] D. Murray and J. Little: Using Real-time Stereo Vision for Mobile Robot Navigation. Auton. Robots, vol. 8, no. 2, pp. 161-171. 2000.

[3]P.H.S Torr and D. W. Murray: Statistical: Detection of Independent Movement from a Moving Camera. In Proc. 3rd British Machine Vision Conference.1992.

[4] E. Vincent and R. Laganiere: Detecting planar homographies in an image pair. Proc. International Symposium on Image and Signal Proceessing and Analysis. 2001

[5] Junghoon Jung, Taekyung Kim, Ohyun Kwon, and Joonki Paik: Plane identification and segmentation from uncalibrated stereo based on evolutionary approach. Intelligent Signal Processing and Communication Systems. 2001

[6] Y. Kanazawa and H. Kawakami: Detection of Planar Regions with Uncalibrated Stereo using Distributions of Feature Points. BMVC. 2004.

[7]B. Boufama and D. O'Connell: Identification and matching of planes in a pair of uncalibrated images, International Journal of Pattern Recognition and Artificial Intelligence, Vol. 17, No. 7 pp. 1127-1143, 2003.

[8] C. Harris and M.J. Stephens: A combined corner and edge detector. Alvey Vision Conference, pages 147–152, 1988

[9] D. Scharstein and R. Szeliski. A taxonomy and evaluation of dense two-frame stereo correspondence algorithms.International Journal of Computer Vision, 47(1/2/3):7-42, April-June 2002

[10] R.Hartley and A Zisserman: Mutiple View Geometry in Computer vision.

[11] http://www.vision.ee.ethz.ch/datasets/index.en.htm

[12] D. Sinclair, A. Blake, Quantitative planar region detection, Int.J. Comput. Vision 18 (1) (1996) 77 – 91

[13] JL Mundy and A. Zisserman: Geometric Invariance in Computer Vision. Vol.First Edtition, MIT Press

[14] http://www.cim.mcgill.ca/~dparks/CornerDetector/index.htm

[15] Yu-Chi Wang, and Pau-Choo Chung: Reconstructing 3D Model of Real Scenes from Photographs, Institute of Computer and Communication National Cheng Kung University, 2005