臺灣博碩士論文加值系統

　　利用多張影像建構環場影像在電腦視覺領域內一直是重要的研究課題，不同於傳統在建立環場接合影像時的諸多限制，例如限制平面場景或攝影機做純旋轉運動，本研究利用流形投影的概念來建構影像接合圖，並且採用贏者更新策略快速區塊比對演算法來進行運動估測，使得本論文發展之即時遠距影像掃描系統得以在任意攝影機運動與任意場景下進行即時環場影像接合。此外，由於長軸畫在數位化的過程中往往十分沒有效率而且繁複，針對此一問題，在本論文裡我們利用所發展之即時遠距影像掃描系統，在等速度與等方向攝影機運動條件限制下分別對長軸畫的不同部份做取像，使得所得到的區域接合影像皆符合Linear Pushbroom Camera Model，並且利用此一模型，找出不同區域接合影像間的幾何對應關係，再用此一關係來將長軸畫做重建。為了達到自動化重建的目的，在本論文中我們將自動找尋與建立區域接合影像間特徵點一對一的點對應關係，以及利用多層解析度顏色融合技術將區域接合影像於不同頻帶內給予不同權重做接合後，再將長軸畫做重建，以減少區域接合影像間色彩的差異在長軸畫重建時所造成的影響。經由許多的實驗結果，証實了本論文提出的方法在長軸畫的重建上具有效率與可行性。

Integrating multiple images taken with perspective cameras for building a panoramic mosaic is an important research topic in computer vision. Several limitations, such as restricted camera motions or scene structures, have been put to be necessary conditions for building panoramic mosaics in some researches. In this thesis, we use the manifold projection model to create panoramic mosaics from videos, and the winner-update strategy for block matching is adopted to speed up the estimation of camera motion. By doing so, we have developed a real-time distance-scanning system for mosaic construction that is suitable for on-line hand-held camera motions and unconstrained scene structures. In addition, the processes for digitalizing wide-field drawings are usually inefficient and complicated. To simplify this problem, we use a real-time distance-scanning system consisting of a video camera moving along a track with constant velocity and constant camera direction to generate a LP-mosaic, which can be modeled as a linear pushbroom camera. In particular, we have conducted mathematical models for the case that multiple linear-pushbroom cameras are used to take a planar scene, which is a relatively unexplored area and has not been well studied in the past. An automatic method for finding matching corners between a pair of LP-mosaic images is adopted, and a multiresolution color blending technique is performed to seamlessly stitch several LP-mosaic images. Good experimental results show the efficiency and feasibility of the proposed method for wide-field drawing reconstruction.

ABSTRACT (in Chinese)/i
ABSTRACT (in English)/ii
ACKNOWLEDGEMENTS (in Chinese)/iii
CONTENTS/iv
LIST OF TABLES/vi
LIST OF FIGURES/vii
Chapter 1　Introduction/1
　1.1　Motivation/2
　1.2　Survey of Related Works/3
　1.3　Background Knowledge/7
　　1.3.1　Manifold Projection Model/7
　　1.3.2　Linear Pushbroom Camera Model/9
　1.4　Thesis Organization/13
Chapter 2　Real-Time Distance-Scanning System/15
　2.1　System Overview/16
　2.2　Image Acquisition and Pre-Processing/18
　2.3　Motion Estimation/19
　　2.3.1　Concept of Winner-Update Strategy/20
　　2.3.2　Winner-Update Algorithm for Block Matching/21
　　2.3.3　Robust Motion Vector Estimation by Using Projected Median/23
　2.4　Image Mosaicing/25
　2.5　Examples of the Conducted Mosaics/26
Chapter 3　Wide-Field Drawing Reconstruction/29
　3.1　Introduction/30
　3.2　Geometric Model for Reconstructing Wide-Field Drawings/30
　　3.2.1　Linear Pushbroom Camera Model for Planar Scene/30
　　3.2.2　Mathematical Models for Reconstructing Wide-Field Drawings/32
　　3.2.3　Exploring Geometry Relationship between LP-Mosaic Image and Reconstructed Wide-Field Drawing/35
　3.3　Finding Matching Points between a Pair of LP-Mosaic Images/40
　　3.3.1　Corner Detection/41
　　3.3.2　Finding Candidate /42
　　3.3.3　Disambiguating Candidate Matches through Relaxation/47
　3.4　Multiresolution Color Blending Technique/50
　　3.4.1　Basic Pyramid Operations/50
　　3.4.2　Reconstructing Seamless Wide-Field Drawings/58
Chapter 4　Experimental Results/61
　4.1　Experimental Results of Real-Time Distance-Scanning System/62
　4.2　Experimental Results of Automatic Wide-Field Drawing Reconstruction System/62
Chapter 5　Conclusions and Future Works/81
　5.1　Conclusions/82
　5.2　Future Works/82
References/85

[1]P. J. Burt and E. H. Adelson, “A Multiresolution Spline with Application to Image Mosaics,” ACM Transactions on Graphics, Vol. 2, No. 4, pp. 217-236, October 1983.
[2]Y. S. Chen, Y. P. Hung, and C. S. Fuh, “Fast Block Matching Algorithm Based on the Winner-Update Strategy,” IEEE Transactions on Image Processing, Vol. 10, pp. 1212-1222, August 2001.
[3]K. W. Chou, A Large-Scale Picture Scanning System Based on Automatic Image Detection and Content Corrections, Master Thesis, Institute of Computer and Information Science, National Chiao Tung University, Hsinchu, Taiwan, June 2000.
[4]R. C. Gonzalez and R. E. Woods, Digital Image Processing, Addison-Wesley Publishing Company, 1992.
[5]C. Guestrin, F. Cozman, and M. G.. Simoes, “Industrial Applications of Image Mosaicing and Stabilization,” In Proceedings of 2nd International Conference on Knowledge-Based Intelligent Electronic System, Vol. 2, pp. 174-183, 1998.
[6]R. Gupta and R. I. Hartley, “Linear Pushbroom Cameras,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 19, No. 9, September 1997.
[7]R. Hartley, Multiple View Geometry in Computer Vision, Cambridge University Press, 2000.
[8]H. J. Ho, Panorama-Guided Visual Tracking and its Application on Augmented Reality, Master Thesis, Department of Computer Science and Information Engineering, National Taiwan University, Taipei, Taiwan, June 2001.
[9]C. T. Hsu and R. A. Beuker, “Multiresolution Feature-Based Image Registration,” In Proceedings of Visual Communications and Image Processing, Vol. 4067 (2000), pp. 1490-1498, Perth, Australia, June 2000.
[10]C. T. Hsu, T. H. Cheng, R. A. Beuker, and J. K. Horng, “Feature-Based Video Mosaic,” In Proceedings of International Conference on Image Processing, Vol.2, pp.887-890, Vancouver, Canada, September 2000.
[11]S. Hsu, H. S. Sawhney, and R. Kumar, “Automated Mosaics Via Topology Inference,” IEEE Computer Graphics and Applications, Vol.22, pp. 44-54, March-April 2002.
[12]C. G. IIarris and M. Stephens, “A Combined Corner and Edge Detector,” In Proceedings of Alvey Conference, pp. 189-192, 1988.
[13]C. G. IIarris, “Determination of Ego-Motion from Matched Points,” In Proceedings of Alvey Conference, 1987.
[14]C. H. Lee and L. H. Chen, “A Fast Motion Estimation Algorithm Based on the Block Sum Pyramid,” IEEE Transactions on Image Processing, Vol. 6, pp. 1587-1591, November 1997.
[15]S. Peleg, B. Rousso, A. Rav-Acha, and A. Zomet, “Mosaicing on Adaptive Manifolds,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 22, No. 10, October 2000.
[16]S. Peleg and J. Herman, “Panoramic Mosaics by Manifold Projection,” IEEE Conference on Computer Vision and Pattern Recognition, pp. 338-343, June 1997.
[17]S. Peleg and J. Herman, “Panoramic Mosaics with VideoBrush,” In Proceedings of DARPA Image Understanding Workshop, pp. 261-264, May 1997.
[18]B. Rousso, S. Peleg, I. Finci, and A. Rav-Acha, “Universal Mosaicing Using Pipe Projection,” In Proceedings of 6th International Conference on Computer Vision, pp. 945-952, January 1998.
[19]R. Szeliski and H. Y. Shum, “Creating Full View Panoramic Image Mosaics and Environment Maps,” In Proceedings of the 24th Annual Conference on Computer Graphics and Interactive Techniques, pp. 251-258, August 1997.
[20]H. Y. Shum and R.Szeliski, “Construction and Refinement of Panoramic Mosaics with Global and Local Alignment,” In Proceedings of 6th International Conference on Computer Vision, pp. 953-956, 1998.
[21]R. Szeliski, “Video Mosaics for Virtual Environments,” IEEE computer Graphics and Applications, Vol.16, pp.22-30, March 1996.
[22]J. A. Noble, “Finding Corners,” Image and Vision Computing, Vol. 6, pp. 121-128, May 1988.
[23]Z. Zhang, R. Deriche, O. Faugeras and Q. T. Luong, ” A Robust Technique for Matching Two Uncalibrated Images Through the Recovery of the Unknown Epipolar Geometry, ” Artificial Intelligence, pp. 87-195, 1995.