proc.dvi

Size: px

Start display at page:

Download "proc.dvi"

さわやまのかみしゃ
5 years ago
Views:

1 TVRSJ Vol.7 No.1, 2002 The Great Buddha Project Λ1 Λ2 Λ1 Λ1 Λ1 Λ3 The Great Buddha Project Digital archive of large-scale cultural heritage Katsushi Ikeuchi Λ1, Ryo Kurazume Λ2, Ko Nishino Λ1, Ryusuke Sagawa Λ1, Takeshi Oishi Λ1, and Yutaka Takase Λ3 Abstract This paper presents an overview of our efforts in modeling the Great Buddha of Kamakura through observation. The Kamakura Great Buddha, built in the 13th Century, is one of the national treasures of Japan; of its height is about 20m, including its pedestal. Currently, the Buddha is standing in the open air. We have started this modeling as the kick-off step of our project to model Japanese cultural heritages by using the modeling-from-reality techniques that make virtual models of real objects through observing the real objects. Due to the size and the circumstance concerning the great Buddha, we have encountered several research issues. For geometric modeling, we have developed a simultaneous registration and parallel volumetric merging algorithms. For photometric modeling, we have developed a new camera-sensor alignment algorithm. We have constructed a CG model of the Main Hall of the Great Buddha, widely briefed its existence from the literature survey, and created a video of the Buddha, measured and constructed with our techniques, in the virtual Main Hall created through the literature survey. We have demonstrated the effectiveness of the algorithms as well as the importance of restoring cultural heritage through these techniques. Keywords : Digital archive, 3D shape modeling, Laser scanner, Photometric modeling 1. (VR) VR "modeling-from-reality (MFR)" (1) (2) (3) *1 *2 *3 ( *1 The University oftokyo *2 Japan Science and Technology Corporation *3 CADCENTER, Inc 3 [27], [34] [12] [20] 2 [22], [23] MFR

Vol.7, No.1, 2002 1 1 Fig. 1 The Great Buddha of Kamakura 13 12.4m 8 8 1264 7 8 13 3 2 1498 1960 [41] 1950 7 2 25mm 12.

2 Vol.7, No.1, Fig. 1 The Great Buddha of Kamakura m [41] mm 12.5mm MFR Stanford [16] IBM [3] Levoy Bernardini Rushmeier 5m 15m MFR CG 2 3 MFR Cyrax 2400( 2) 100m mm

: The Great Buddha Project Cyrax 2400 1 3 4 24 1

Range 100[m] Accuracy 6 [mm] Resolution 0.

3 : The Great Buddha Project Cyrax Cyrax 2400 Table 1 Specification of Cyrax 2400 Range 100[m] Accuracy 6 [mm] Resolution 0.25 [mm] at 50[m] N. of points Cyrax 2400 Fig. 2 Cyrax Fig. 3 Scanning the Great Buddha 4 Fig. 4 Some of range data obtained [25] 2 [1], [2], [5], [6], [10], [18], [19], [21] : Euclidian [14] closed form 3 Besl Kay ICP [4] [18] [6], [19]

4 Vol.7, No.1, 2002 [1], [5], [19] Cyrax for(each model)f for(eachmodel point)f for(each scene)f scene point = findpointmate(modelpoint); dist += computedist (model point, scene point); g g g step[model] = computestep(dist); for(each model)f transform(model,step[model]); g "model" "scene" "step" 1 computestep M [27] M ComputeStep E(P )= 1 N X i2n ρ(z i (P )) (1) P model t 4 q P =(t; q) T N z i (P ) i 3 ρ E(P ) E(P ) ) i (P ) =0 w(z i ) w(z i )= (3) z ) X i2n w(z i )z i (P (4) random sample consensus [29] w(z i ) w(z) = 1+ 1 z 2 1 (5) 2 ff z i z i (P )=j R(q)x i + t y i j (6) R(q) P t x i y i model scene 4 z i i (P ) ψ 2(x i + t y) 4x i (t y)! (7)

: The Great Buddha Project 1 5 5 Fig. 5 Aligned results 2.

8 marching-cubes algorithm[17] [27] Dual Pentium III 800MHz 1GB 8 PC CPU

5 : The Great Buddha Project Fig. 5 Aligned results consensus [27] , marching-cubes algorithm[17] [27] Dual Pentium III 800MHz 1GB 8 PC CPU 16 PC PC PC 2 2. (2-1) PC (2-2) PC 6 1,2,3 PC1,2,3 x 1 PC1 2 PC2 PC Fig. 6 6 Parallel computation of signed distance

6 Vol.7, No.1, PC PC PC 3 3. (3-1) (3-1) PC PC 2, Fig. 8 Merged result: The model contains 3 millions points and 5.5 millions triangles Surface Octree PC1 PC2 PC3 PC4 PC1 PC2 PC3,... 7 Fig. 7 Parallel traversal of partial trees PC PC 1 PC PC 8 PC 2 CPU PC OGIS Cyberwares [24]

7 : The Great Buddha Project Viola [32] Allen [31] [33], [34], [35] albedo ERIM Perceptron Cyrax Canny [39] Canny D -2D 9(a) Canny 3D 9(b) 2D (a) Reflectance image (b) Color texture image 9 Fig. 9 Reflectance and color images 2 M

8 Vol.7, No.1, Fig. 12 Aligned color texture on the whole body of the 3D geometric model Fig. 13 Residue of gold leaf 10 Fig. 10 Aligned intensity edges with reflectance edges 14 Fig. 14 Golden Buddha 11 Fig. 11 Aligned color texture on the 3D geometric model

: The Great Buddha Project 3D CAD 30 CG CG 17 15 Fig.

16 16 Drawings of the Jodo-do (Jodo-ji) 5. 3 1. 2. 3. 1. 2. 3. - CREST [1] R. Benjemma and F. Schmitt.

On Recent Advances in 3-D Digital Imaging and Modeling, pp 113-120, 17 Fig.

Towards a general multi-view registration technique. IEEE Trans. Patt. Anal. Machine Intell., 18(5)540-547, May. 1996. [3] F.

A method for registration of 3-D shapes. IEEE Trans. Patt. Anal. Machine Intell., 14(2):239-256, Feb. 1992. [5] G. Blais and M.D. Levine.

[6] Y. Chen and G. Medioni. Object modeling by registration of multiple range images.

9 : The Great Buddha Project 3D CAD 30 CG CG Fig. 15 Drawings of the Main Hall, Todai-ji, reconstructed in the 13th century (by Minoru Oka) Fig Drawings of the Jodo-do (Jodo-ji) CREST [1] R. Benjemma and F. Schmitt. Fast global registration of 3D sampled surfaces using a multi-z-buffer technique. In Proc. Int. Conf. On Recent Advances in 3-D Digital Imaging and Modeling, pp , 17 Fig. 17 Golden Buddha and the Main Hall in the 13th century May [2] R. Bergevin, M. Soucy, H. Gagnon and D. Laurendeau. Towards a general multi-view registration technique. IEEE Trans. Patt. Anal. Machine Intell., 18(5) , May [3] F.Bernardini and H.Rushmeier, "The 3D Model Acquisition Pipeline" Eurographics [4] P.J. Besl and N. D. McKay. A method for registration of 3-D shapes. IEEE Trans. Patt. Anal. Machine Intell., 14(2): , Feb [5] G. Blais and M.D. Levine. Registering multiview range data to create 3D computer objects. IEEE Trans. Patt. Anal. Machine Intell., 17(8): , Aug [6] Y. Chen and G. Medioni. Object modeling by registration of multiple range images. Image and Vision Computing, 10(3): , Apr [7] B. Curless and M. Levoy A volumetric method for building complex models from range images. In Proc. SIGGRAPH 96, pp [8] C. Dorai, G. Wang, A.K. Jain and C. Mercer. From images to models: Automatic 3D object model construction from multiple views. In Proc. IAPR ICPR, pp , [9] J.H. Friedman, J.L. Bentley and R.A. Finkel. An

10 Vol.7, No.1, 2002 algorithm for finding best matches in logarithmic expected time. ACM Trans. On Mathematical Software, 3(3): , [10] G. Godin, M. Rioux and R. Baribeau. Threedimensional registration using range and intensity information. In Proc. SPIE vol.2350: Videometrics III, pp , [11] H. Hoppe, T. DeRose, T. Duchamp, J. McDonald, and W. Stuetzle, "Surface reconstruction from unorganized points," Proc. SIGGRAPH 92, pp [12] Y. Sato, M. D. Wheeler, and K. Ikeuchi, "Object shape and reflectance modeling from observation", Proceedings of ACM SIGGRAPH 97, In Computer Graphics Proceedings, Annual Conference Series 1997, ACM SIGGRAPH, pp , August [13] H. Jin, T. Duchamp, H. Hoppe, J.A. McDonald, K. Pulli and W. Stuetzle. Surface reconstruction from misregistered data. In Proc. SPIE vol.2573: Vision Geometry IV, pp , [14] A. Johnson and M. Hebert. Surface registration by matching oriented points. In Proc. Int. Conf. on Recent Advances in 3-D Digital Imaging and Modeling, pp , May [15] A. Johnson and S.B. Kang. Registration and integration of textured 3-D data. In Proc. Int. Conf. On Recent Advances in 3-D Digital Imaging and Modeling, pp , May [16] M. Levoy et. al. The Digital Michelangelo Project. In Proc. SIGGRAPH 00, pp [17] W.E. Lorensen, W.E. and H.E. Cline. Marching Cubes: a high resolution 3D surface reconstruction algorithm. In Proc. SIGGRAPH96, pp [18] T. Masuda, K. Sakaue and N. Yokoya. Registration and integration of multiple range images for 3-D models construction. In Proc. CVPR, pp , Jun [19] P. Neugebauer. Geometrical cloning of 3D objects via simultaneous registration of multiple range images. In Proc. Int. Conf. on Shape Modeling and Application, pp , Mar [20] K.Nishino, Y.Sato and K.Ikeuchi, "Eigen-Texture Method: Appearance Compression based on 3D Model", in Proc. of Computer Vision and Pattern Recognition '99, vol.1, pp , Jun., [21] K. Pulli. Multiview registration for large data sets. In Proc. of Second Int. Conf. on 3D Digital Imaging and Modeling, pp , Oct [22] I. Sato, Y. Sato, and K. Ikeuchi, "Acquiring a radiance distribution to superimpose virtual objects onto a real scene," IEEE Trans Visualization and Computer Graphics, Vol. 5, No. 1, pp.1-12, January [23] I. Sato, Y. Sato, and K. Ikeuchi, "Illumination distribution from brightness in shadows: adaptive estimation of illumination distribution with unknown reflectance properties in shadow regions," Proceedings of IEEE ICCV'99, pp , September [24] R. Tsai. A Versatile Camera Calibration Technique for High Accuracy Machine Vision Metrology Using Off-the-Shelf TV Cameras and Lenses. IEEE J. R&A, 3( 4,): , [25] G. Turk and M. Levoy. Zippered polygon meshes from range images. In Proc. of SIGGRAPH 94, pp , Jul [26] P. Viola and W.M. Wells III. Alignment by maximization of mutual information, International Journal of Computer Vision, 24(2): , [27] M. Wheeler, Y. Sato, and K. Ikeuchi, Consensus surfacs for modeling 3D objects from multiple range images, Proc. ICCV pp , Jan, [28] Z. Zhang. Iterative point matching for registration of free form curves and surfaces. International Journal of Computer Vision, 12(2): , [29] M.A. Fischler and R.C. Bolles,Random sample consensus: A paradigm for model fitting with applications to image analysis and automated cartography, Communications of the ACM, Vol.16, No.24, [30] Mark D. Elstrom and Philip W. Smith,Stereo- Based Registration of Multi-Sensor Imagery for Enhanced Visualization of Remote Environments,Proceedings of the 1999 IEEE International Conference on Robotics & Automation, pp , [31] Ioannis Stamos and Peter K. Allen, Integration of Range and Image Sensing for Photorealistic 3D Modeling, Proceedings of the 2000 IEEE International Conference on Robotics and Automation, pp , [32] P. Viola and W.M. Wells III, Alignment by maximization of mutual information, International Journal of Computer Vision, Vol.24, No.2, pp , [33] M. D. Wheeler, "Automatic Modeling and Localization for Object Recognition", Technical Report (Ph.D. Thesis), CMU-CS , School of Computer Science, Carnegie Mellon University, October, [34] M. D. Wheeler and Katsushi Ikeuchi, "Sensor Modeling, Probabilistic Hypothesis Generation, and Robust Localization for Object Recognition", IEEE Transactions on Pattern Analysis and Machine Intelligence,Vol. 17, No. 3, March [35] D-II Vol J83-D-II No.2 pp [36] P. Debevec, D.J. Taylor, and J. Malik, Modeling and rendering architecture from photographs; A hybrid geometry and image-base approach, Proc. of SIGGRAPH'96, pp.11-20, [37] P. Debevec, Y. Yu, and G. Borshukov, Efficient view-dependent image-based rendering with projective texture-mapping 9th Eurographics workshop on rendering, pp , [38] Daisuke Miyazaki, Takeshi Ooishi, Taku Nishikawa, Ryusuke Sagawa, Ko Nishino, Takashi Tomomatsu, Yutaka Takase, Katsushi Ikeuchi,The Great Buddha Project: Modelling Cultural Heritage through Observation,VSMM2000 (6th international conference on virtual systems and multimedia), pp , [39] J. F. Canny, A computational approach to edge detection, IEEE T rans. Pattern Analysis and Machine Intelligence, Vol.8, No.6, [40] [41] - -,Vol.12, No.6,pp ,1960 ( )

11 : The Great Buddha Project MIT CMU 1996 (ICCV-90)CVPR-91 AIJ IEEE R&A -98 OSA IEEE(Fellow) MIT CG CG/VR M.S.Arch.S MIT VSMM ( )

proc.dvi

proc.dvi The Great Buddha Project Λ1 Λ2 Λ2 Λ2 Λ3 Λ2 Λ1 Λ2 Λ3 ( (VR) VR (1) (2) (3) modeling-from-reality (MFR) The Great Buddha Project Digital archive of large-scale cultural heritage Ryo Kurazume Λ1 Ko Nishino