The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B5. Beijing 2008 
1103 
was signalised with a 7 x 7 cm, mostly black and white, 
adhevise target stuck on PVC. For the walls, due to difficulties 
in placing signalised points there, only a few signalised points 
were used in accessible areas. Further control and check points 
were well-defined natural features measured with a total station 
and points transferred from the laser scanning intensity 
information (see Section 2.2). 
Figure 4. Part of the point cloud of the interior east wall showing 
the Belvedere at the NE side of the Acropolis wall. 
Special targets such as spheres and adhesive objects were used 
in the scans in order to co-register partial scans to a common 
system. For example, the scanning of a particular sector of the 
wall took approximately one hour, giving in total 2 million 3D 
points. This is in agreement with the project specifications, 
whereby the resolution of the point cloud had to be 10mm for 
the walls. 
As a separate task, the laser scanning of Erechtheion was 
Figure 2. The hill of Acropolis (aerial photo). In the center carried out by two other scanners (Surphaser 25HSX and Leica 
the Parthenon, left the Propylaia, at the top Erechtheion, on HD3000). This approach is described in more details in a 
the right the old Acropolis museum. different paper (El-Hakim et al., 2008). 
2.2 Laser scanning 
For the point cloud collection, the Trimble GX 3D scanner 
(time-of-flight) was deployed. The range of the instrument 
extends up to 300m, with an accuracy of 2.5 mm at 100m 
distance, and measuring speed of 5000 points/sec. A total of 
146 scans were collected from 53 different scanning positions, 
covering the wall façades (inner and outer sides) and the 
surrounding area, including the rock. In Figures 3 and 4, two 
point clouds are shown of the south and east wall respectively. 
The points are textured with the intensity map from the laser 
scanner allowing the identification of specific components of 
the monument. This information gives additional image- 
identifiable ground measurements on the object, which can be 
treated as control points during the bundle adjustment of the 
wall images from the digital camera. 
3. PROCESSING 
3.1 Camera calibration 
The internal geometry of the camera was recovered through 
bundle adjustment with the Leica LPS software using a network 
of 29 control points measured with accuracy less than 2mm. 
Camera 
parameters 
A posteriori o 0 =0.3069 pixel 
c (mm) 
x 0 (mm) 
yo (mm) 
ki 
k 2 
46.0998 ±0.0136 
0.0393 ±0.0063 
0.0424 ±0.0063 
-4.0207000e-05 ±3.9364e-007 
1,9295000e-08 ±4.9547e-010 
Table 1. Calibration results (parameter values and standard 
deviations) for the 5m distance imagery (walls). 
Camera 
parameters 
A posteriori o 0 =0.2867 pixel 
c (mm) 
x 0 (mm) 
yo (mm) 
ki 
k 2 
45.7830 ±0.0110 
0.0535 ±0.0074 
0.0347 ±0.0074 
-4.0700000e-05 ±3.8924e-007 
2.0735000e-08 ±4.4342e-010 
Table 2. Calibration results (parameter values and standard 
deviations) for the 22m distance imagery (top view). 
Figure 3. Part of the point cloud in the south wall. 
Two adjustments were performed, one at a distance of 5m 
concerning the images of the walls and a second one using