3D Acquisition of Archaeological Heritage from Images 
387 
The following example was recorded at Sagalassos in Turkey, where footage of the ruins of an ancient fountain was taken. The 
fountain video sequence consists ot 250 frames. A large part of the original monument is missing. Based on results of archaeological 
excavations and architectural studies, it was possible to generate a virtual copy of the missing part. Using the proposed approach the 
virtual reconstruction could be placed back on the remains of the original monument, at least in the recorded video sequence. The top 
part of Figure 8 shows a top view of the recovered structure before and after bundle-adjustment. Besides the larger reconstruction 
error it can also be noticed that the non-refined structure is slightly bent. This effect mostly comes from not taking the radial 
distortion into account in the initial structure recovery. In the rest of Figure 8 some frames of the augmented video are shown. 
6. CONCLUSION 
In this paper an approach for obtaining virtual models with a hand-held camera was presented. The approach utilizes different 
components that gradually retrieve all the information that is necessary to construct virtual models from images. Automatically 
extracted features are tracked or matched between consecutive views and multi-view relations are robustly computed. Based on this 
the projective structure and motion is determined and subsequently upgraded to metric through self-calibration. Bundle-adjustment 
is used to refine the results. Then, image pairs are rectified and matched using a stereo algorithm and dense and accurate depth maps 
are obtained by combining measurements of multiple pairs. From these results virtual models can be obtained or, inversely, virtual 
models can be inserted in the original video. 
This technique was successfully applied to the acquisition of virtual models of archaeological sites. There are multiple advantages: 
the on-site acquisition time is restricted, the construction of the models is automatic and the generated models are realistic. The 
technique allows some more promising applications like 3D stratigraphy, the (automatic) generation and verification of building 
hypothesis and mixing archaeological remains with virtual reconstruction in video. 
ACKNOWLEDGEMENT 
Marc Pollefeys and Kurt Cornelis are respectively post-doctoral fellow and research assistant of the Fund for Scientific Research - 
Flanders (Belgium). The financial support of the FWO project G.0223.01, the ITEA BEYOND of the IWT and the 1ST-1999-20273 
project 3DMURALE are also gratefully acknowledged. 
REFERENCES 
P. Beardsley, A. Zisserman and D. Murray, 1997. Sequential Updating of Projective and Affine Structure from Motion, International 
Journal of Computer Vision (23), No. 3, pp. 235-259. 
Y. Chen and G. Medioni, Object modelling by registration of multiple range images, Image and Vision Computing, vol. 10, no. 3, pp. 
145-155, April 1992. 
K. Cornelis, M. Pollefeys, M. Vergauwen and L. Van Gool, 2001. Augmented Reality from Uncalibrated Video Sequences, In M. 
Pollefeys, L. Van Gool, A. Zisserman, A. Fitzgibbon (Eds.), 3D Structure from Images - SMILE 2000, Lecture Notes in 
Computer Science, Vol. 2018, Springer-Verlag. pp. 150-167. 
I. Cox, S. Hingorani and S. Rao, 1996. A Maximum Likelihood Stereo Algorithm, Computer Vision and Image Understanding, Vol. 
63, No. 3. 
B. Curless and M. Levoy, 1996. A Volumetric Method for Building Complex Models from Range Images. Proc. SIGGRAPH. pp. 
303-312. 
P. Debevec, C. Taylor and J. Malik. 1996. Modelling and Rendering Architecture from Photographs: A Hybrid Geometry- and 
Image-Based Approach. Proc. SIGGRAPH. pp. 11-20. 
L. Falkenhagen, 1997. Hierarchical Block-Based Disparity Estimation Considering Neighbourhood Constraints, Proc. International 
Workshop on SNHC and 3D Imaging, Rhodes, Greece, pp.l 15-122. 
O. Faugeras, 1992. What can be seen in three dimensions with an uncalibrated stereo rig, Computer Vision - ECCV'92, Lecture 
Notes in Computer Science, Vol. 588, Springer-Verlag, pp.563-578. 
M. Fischler and R. Bolles, 1981. RANdom SAmpling Consensus: a paradigm for model fitting with application to image analysis and 
automated cartography, Commun. Assoc. Comp. Mach., 24:381-95. 
C. Harris and M. Stephens, 1988. A combined corner and edge detector, Fourth Alvey Vision Conference, pp. 147-151. 
R- Hartley, R. Gupta, and T. Chang, 1992. Stereo from uncalibrated cameras. Proc. Conference Computer Vision and Pattern 
Recognition, pp. 761-764. 
R- Koch, 1996. Automatische Oberflachenmodellierung starrer dreidimensionaler Objekte aus stereoskopischen Rundum-Ansichten, 
PhD thesis, Univ. of Hannover, Germany.