10DOSCOpic or 
1. It should be 
orientation to 
aptured, while 
with in-situ 
Structure in a 
hole elevation 
in conjunction 
d parts, would 
tor was used 
density and 
o the complex 
|. 
  
t area. 
aser scanning 
plotting of the 
palm receives 
rated optical 
ality of Cyrax 
| the reduction 
applications. 
he decision to 
pe alterations, 
isolated areas 
very difficult 
surveying and 
r to cover the 
tance did not 
of the model 
  
Figure 7. Point Cloud 
The point cloud was determined in proportion to the object's 
detail. For instance, for the orchestra and the stages near the 
orchestra (the most characteristic area of the monument) the 
point cloud was 2-3cm, 0.5cm for the relief restoration (statues) 
and -15cm for the rest area. The integration of the scanning 
time was depended on the size of the corresponding scanning 
area and the point cloud, and it was about 10 min, while the 
whole Theatre's area was scanned in one day. 
The results were impressive, while the point cloud was too 
dense, that it gave the impression of a photographic image. The 
3D MESH that was automatic calculated was really excellent. 
2.6 CYRAX 2500 measuring procedure 
The Cyrax was put across the field of view. A camera build-in 
the system beams a snapshot of the observation area in the 
monitor of the computer. The user selects the part that wishes to 
study and determines the point cloud in a specific range that the 
system could direct measure (i.e. at 45.2m the point cloud is 
lem). The system automatically calculates the final number of 
points that would be measured in columns and rows. 
During taking measurements, there is a possibility for real-time 
tour (pan, zoom, rotate) in the model, through the use of the 
laptop, in order to spot “shades” (areas that have not been 
scanned as they area located behind others) and its next 
coverage from another view. 
The system was not oriented correspondingly to the object, and 
it computed the model in an arbitrary reference system, which 
was different for each scanning. Thus, the registration 
procedure was followed. For that reason, special targets were 
detected by the system during scanning and their location 
(according to the model) was estimated with high accuracy. The 
coordinates of the targets were included to the united reference 
System, so that the model got standard coordinates through 
registration. Three (3) such targets were required for each 
scanning. The whole registration was performed by the Cyclone 
software that accompanies the Laser. 
2.7 Point cloud 
The final product of each scanning was not a common model 
(tin, grid), but a point cloud without a constant distance 
between them, having colour information for the material that 
represented. The point cloud behaved as the common models 
(change of viewing angle, turn, transfer, etc.) and it could be 
transformed to other more common models, which were ease in 
the use of format management, like tin and mesh, through the 
Cyclone software. Moreover, it could be exported and imported 
foa modelling program for further processing. 
The Theatre's model was created by the conjunction of twenty 
(20) different scans using Cyclone. Furthermore, the Theatre's 
sections were estimated at different directions. Next, the point 
cloud was converted to ASCII format. It should be noted that 
the point cloud was about 12cm. 
  
Figure 8. 3D-mesh 
A. Visual Basic program was used for reducing the file size at 
half and to one third so that the new files could be better 
handled. The new files were imported in Helava DPW770, and 
converted to a DTM with a grid size of 5-10cm, in order to 
control and compare these results with those from the 
photogrammetric methodology. 
3. CONCLUSIONS 
The combination between the Cyrax results and those of 
photogrammetry and the terrestrial geodetic observations gave 
amazing results. 
The differences in X, Y, Z on specific points were of the order 
of 1-2cm. Thus, it is difficult to estimate if the orientation error 
was due to the system or the accuracy of the ground-based 
observation. Moreover, the entire point coverage was quite 
difficult to be obtained and almost impossible using ground- 
based observation. The accuracy and the detail were so high 
that it was difficult to identify elements that were documented 
in the field. For example, the elevation differences in the 
Orchestra's stones (Figure 9) were less than 9mm. 
  
Figure 9. Details of the Orchestra. 
In any case, the information management was really difficult. 
The files were enormous in size and each processing was time- 
consuming, requiring the maximum of the computer's 
s