近年來由於電腦網路已發展的相當成熟，傳輸速度也達到一般使用者的需求，因此
許多公司、政府機構的文件資料也不再限定於紙本的形式，電子化資料變成一種共同的
傳媒。通常他們將資料數位化後放在網路上供人查詢、瀏覽，所以掃瞄器的功能對許多
人更顯重要，然而此裝置在使用上不夠便利且相當佔空間，因此本論文擬採用一般視訊
攝影機來進行掃瞄的工作，期望能提高使用上的便利性。首先對文件影像進行歪斜復
原，再對復原後的重疊影像進行接合；整個文件影像接合的系統共分為三個主要程序。
在前處理的程序中，我們先將所擷取到的文件影像轉為灰階影像，接著對它進行二值
化。由於一般單一門檻值的二值化方法對於陰影的部分較難處理，因此我們利用改良式
的Niblack 法對文件影像進行二值化，它可有效改善陰影對二值化的影響，然後利用相
連區塊偵測演算法來過濾文件影像中的非文字資訊，經實驗展示此方法可過濾大多數的
非文字資訊，而提高了估測歪斜角度的穩定性。在歪斜復原程序中，我們採用分析投影
圖的方式來偵測文件影像的歪斜角度，由於它必須對影像不斷地旋轉和投影，使得整個
偵測的步驟相當耗時，因此我們改用階層式搜尋法的概念，先求出粗略的歪斜角度再逐
步逼近找出一精確的歪斜角度。而在文件影像接合的程序，我們採用無特徵點的接圖法
。由於我們先對文件影像進行歪斜復原，所以只需計算平移參數及縮放參數。重疊影像
的接合係利用子區塊的比對法而得，因此可節省區塊比對所花費的時間。截至目前的實
驗結果顯示：我們所提的方法能夠正確並且快速地復原歪斜文件影像及接合多張復原後
的重疊影像。

In recent years, computer networks have been well developed, so that the transmission speeds can achieve the common user's demand. Therefore, many company and government documents are no longer confined to the form of paper; the digital information turns into one kind of popular communication media. Usually, people put digital information in the network to provide persons for inquiring or browsing the documents. Thus, the scanner is very important for a lot of people. However, this device is not convenient for use and occupies the
space too large. In this thesis, we plan to employ a general web camera to carry on the work of a scanner. We expect it can enhance the convenience in the use. First, we recover skewed document images; secondly, we register the overlapped recovered document images. There are three main procedures in our implemented document image registration system. In the preprocessing procedure, the original document image which we capture is transferred into a grey-scale one and then binarized. Single threshold binarization methods are usually very difficult to remove the shadows in the images. An improved Niblack’s algorithm is used to binarize the grey-scale image, which may effectively reduce the influences caused by shadows. Subsequently, a connected component detection algorithm is applied to filtering non-textual information in the document image. The experimental results demonstrate this algorithm can filter the majority of the non-textual information to raise the stability of estimating skew angles. In the skew recovery procedure, a projection profile analysis method is adopted to detect the skew angle of a document image. Because this method must rotate and project the image degree by degree, the entire of detecting skew angles is quite time-consuming. To overcome this problem, we use a hierarchical search method to detect a coarse skew angle and then to approximately refine it. In the document image registration procedure, we apply a featureless image registration method. Since the skewed document image has been recovered, only translation and scale parameters need to compute between
two overlapped recovered images. The registration method is based on sub-block matching. It can spend less computation time resulting from the hierarchical block matching concept. So far, our proposed methods can recovery skewed document images and register multiple overlapped recovered images correctly and fluently.

[1] J. J. Hull, “Document image skew detection: survey and annotated bibliography,“ J. J.
Hull and S. L. Taylor(eds.) Document Analysis Systems II, World Scientific, pp.
40-64, 1998.
[2] D. S. Bloomberg, G. E. Kopec, L. Dasari, “Measuring document image skew and
orientation,” L. M. Vincent and H.S. Baird(eds.) in Proc. of SPIE: Document
Recognition II, San Jose, California, vol. 2422, pp. 302-316, 1995.
[3] H. S. Baird, “The skew angle of printed documents,” L. O’Gorman and R. Kasturi
(eds.) Document Image Analysis, pp. 204-208, 1995.
[4] N. Liolios, N. Fakotakis, G. Kokkinakis, “On the generalization of the form
identification and skew detection problem,” Pattern Recognition, vol. 35, pp. 253-264,
2002.
[5] A. Hashizume, P. S. Yeh, A. Rosenfeld, “A method of detecting the orientation of
aligned components,” Pattern Recognition Letters, vol. 4, pp. 125-132, 1986.
[6] L. O’Gorman, “The document spectrum for page layout analysis,” IEEE Trans. on
Pattern Analysis and Machine Intelligence, vol. 15, no. 11, pp. 1162-1173, 1993.
[7] X. Jiang, H. Bunke, D. Widmer-Kljajo, “Skew detection of document images by
focused nearest-neighbor clustering,” in Proc. 5th Int. Conf. on Document Analysis
and Recognition, Bangalore, pp. 629-632, 1999.
[8] N. Liolios, N. Fakotkis, G. Kokkinakis, “Improved document skew detection based on
text line connected component clustering,” in Proc. of Int. Conf. on Image Processing,
Thessaloniki, vol. 1, pp. 1098-1101, 2001.
[9] S. N. Srihari, V. Govindraju, “Analysis of textual image using the Hough transform,”
Machine Vision Applications, vol. 2, pp. 141-153, 1989.
[10] H. F. Jiang, C. C. Han, K. C. Fan, “A fast approach to the detection and correction of
skew documents,” Pattern Recognition Letter, vol. 18, pp. 675-686, 1997.
[11] A. Amin, S. Fischer, “A document skew detection method using the Hough
transform,” Pattern Analysis and Applications, vol. 3, no. 3, pp. 243-253, 2000.
[12] U. Pal, B. B. Chaudhuri, “An improved document skew angle estimation technique,”
Pattern Recognition Letter, vol. 17, pp. 899-904, 1996.
[13] A. Rundle, “Optimum scan angle determining means,” International Business
Machines Inc., U.S. Patent 3,831,146, August 20, 1974.
[14] L. Dasari, D. S. Bloomberg, “Rapid detection of page orientation,” Xerox
Corporation, U.S. Patent 5276742, January 4, 1994.
[15] W. Niblack, An Introduction to Digital Image Processing, Prentice-Hall, Englewood
Cliffs, New Jersey, pp. 115-116, 1986.
[16] Zheng Zhang, Chew Lim Tan, “Recovery of distorted document images from bound
volumes,” in Proc. 6th Int. Conf. on Document Analysis and Recognition, pp.
429-433, 2001.
[17] Barbara Zitova, Jan Flusser, “Image registration methods: a survey,” Image and
Vision Computing, vol. 21, pp. 977-1000, 2003.
[18] W. K. Pratt, Digital Image Processing, 2nd Edition, Wiley, New York, 1991.
[19] A. Goshtasby, G.C. Stockman, “Point pattern matching using convex hull edges,”
IEEE Trans. on Systems, Man and Cybernetics, vol. 15, pp. 631–637, 1985.
[20] M. Roux, “Automatic registration of SPOT images and digitized maps,” in Proc. of
the IEEE Int. Conf. on Image Processing ICIP’96, Lausanne, Switzerland, pp.
625-628, 1996.
[21] Y.C. Hsieh, D.M. McKeown, F.P. Perlant, “Performance evaluation of scene
registration and stereo matching for cartographic feature extraction,” IEEE Trans.
on Pattern Analysis and Machine Intelligence, vol. 14, pp. 214–237, 1992.
[22] S. Moss, E.R. Hancock, “Multiple line-template matching with EM algorithm,”
Pattern Recognition Letters, vol. 18, pp. 1283–1292, 1997.
[23] X. Dai, S. Khorram, “Development of a feature-based approach to automated image
registration for multitemporal and multisensor remotely sensed imagery,” in Int.
Geoscience and Remote Sensing Symposium IGARSS’97, Singapore, pp. 243–245,
1997.
[24] H. Maitre, Y. Wu, “Improving dynamic programming to solve image registration,”
Pattern Recognition, vol. 20, pp. 443–462, 1987.
[25] S.Z. Li, J. Kittler, M. Petrou, “Matching and recognition of road networks from
aerial images,” in Proc. 2nd European Conf. on Computer Vision ECCV’92, Italy,
pp. 857–861, 1992.
[26] G. Stockman, S. Kopstein, S. Benett, “Matching images to models for registration
and object detection via clustering,” IEEE Trans. on Pattern Analysis and Machine
Intelligence, vol. 4, pp. 229–241, 1982.
[27] M. Ehlers, “Region-based matching for image registration in remote sensing
databases,” in Proc. International Geosciences and Remote Sensing Symposium
IGARSS’91, Finland, pp. 2231-2234, 1991.
[28] D. Bhattacharya, S. Sinha, “Invariance of stereo images via theory of complex
moments,” Pattern Recognition, vol. 30, pp. 1373–1386, 1997.
[29] L.M.G. Fonseca, B.S. Manjunath, “Registration techniques for multisensor remotely
sensed imagery,” Photogrammetric Engineering and Remote Sensing, vol. 62, pp.
1049–1056, 1996.
[30] R. J. Althof, M. G. J. Wind, J. T. Dobbins, “A rapid and automatic image
registration algorithm with subpixel accuracy,” IEEE Trans. on Medical Imaging,
vol. 16, pp. 308–316, 1997.
[31] D. I. Barnea, H. F. Silverman, “A class of algorithms for fast digital image
registration,” IEEE Trans. on Computing, vol. 21, pp. 179–186, 1972.