The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B5. Beijing 2008 
4. RESULTS AND ACCURACY ANALYSIS 
4.1 The First Model 
The Project first phase calls for the creation of a lower 
resolution model (10 million polygons, 15mm lateral spacing 
and more on flat surfaces) that is complete and interactively 
viewable on a high-end PC. The objective is to make sure that 
all the monument has been covered and that the data can be 
processed and integrated successfully. Figure 7 shows results of 
this phase. Notice the lack of fine details on the Caryatids. 
I! 
sHiftl 
Figure 8. High-resolution model of a Caryatid column. 
Work on the full resolution model is in progress since merging 
and repairing such huge data sets remains time consuming. 
Some parts have been completed. Figure 8 shows the full 
resolution model of one of the Caryatids columns. Although the 
original points were captured at 1-2 mm lateral spacing, 
simplifications were necessary to reduce data size. On sharp 
features, the original data was not reduced, while on flat areas 
reductions were made without sacrificing details. 
4.2 Initial Accuracy Test Results 
Regions of the Erechtheion, with variable surface roughness 
and marble age, were selected for testing the performance of the 
laser scanners on this material. One test was to estimate the 
amount of apparent laser penetration on marble. A sheet of 
paper was placed on a nearly flat wall of the monument, and 
then scanned along with the surrounding region. A plane is 
fitted to the paper surface and the normal distance between 
every scanned point and that plane was computed. 
As shown in figure 9 the colour-coded depth indicates that the 
marble surface is about 5mm lower than paper surface, which is 
not true since the paper was right on the marble surface (paper 
thickness is about 0.1 mm). This systematic range error may be 
attributed to a combination of laser penetration and unusual 
backscattering properties of the laser light on this type of 
marble. The spatial and temporal distribution of the laser light 
collected by the scanner seems to affect adversely the time 
delay estimate of the modulated signal. One also observes that 
adhesive lines between the marble blocks look as if they were 
higher than the marble surface, while in reality they are lower. 
Both scanners exhibited a similar behaviour, with the Leica® 
scanner showing less penetration. The above tests prove that 
there is an apparent significant laser penetration into the marble, 
however the exact amount of that penetration and how to 
correct it over marbles of different age are still not entirely 
solved problems. 
Figure 9. Marble penetration: comparison with paper sheet. 
Green is zero, red is 7mm positive, and purple is 7mm negative. 
5. CONCLUSIONS 
The data acquisition, processing, problems encountered, and 
some results and analysis of the digital reconstruction of the 
Erechtheion have been presented. Multiple sensors and 
techniques were employed and some tools had to be adapted for 
use with large complex monuments. Such tools, along with the 
gained experience and lessons learned will be valuable for 
future projects of modelling comparable structures. Full model 
at compressed resolution (10 million polygons) has been 
completed, while work on the full resolution model (several 
billions of polygons) is currently in progress. An effective and 
precise solution to the problems of laser scanning marble 
surfaces is being developed. Additional future work includes 
the integration of 3D data of landscapes and models of missing 
top parts of the monument produced by the balloon-based aerial 
images, and the final texturing and lighting of the full model 
under different times of the day and different seasons. A high 
quality computer animation of the site is also in production. The 
incorporation of the 3D model in a GIS database with other 
information will be the final phase of the project. 
ACKNOWLEDGEMENTS 
This project is financed by the EU and the Government of 
Greece and supervised by the Acropolis Restoration Service 
(YSMA), Hellenic Ministry of Culture. Our partners included 
Elliniki Photogrammetriki Ltd. (Elpho), Athens, Swiss Federal 
Institute of Technology - ETH-Zurich, and Geotech O.E., 
Athens. Leica Geosystems, Switzerland and Basis Software Inc., 
USA, supplied the laser scanners. 
REFERENCES 
Aliaga, D., Cohen, J., Wilson, A., Baker, E., Zhang, H., Erikson, 
C., Hoff, K., Hudson, T., Sturzlinger, W., Bastos, R., Whitton, 
M., Brooks, F. Manocha, D., 1999. MMR: An interactive 
massive model rendering system using geometric and image- 
based acceleration. In Symposium on Interactive 3D Graphics, 
Atlanta, USA, 26 April -2 May, pp. 199-206. 
1081