bul 2004
|
|
|
|
|
|
|
aluation. |
amera at |
|
|
|
|
|
|
|
or of 13
hown in
ad been
tation
ctively.
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004
Figure 3. 3D design of the object.
3.2 Laser Scanning Method:
The study area had been scanned by using local coordinate
system. During scanning process with control points each
different parts side of building had been located for registration
process. Cyrax HSD2500 sensitive specific targets had been
used. The three side of building was scanned with 6 mm point
diameter. All scanned sides of building registered with 1,4 mm.
accuracy and 3D point cloud model of the building was
generated (Figure 4).
Figure 4. Point clouds of whole building
3D modelling of the laser scanned study area has been done by
using Cyclone software. Single patches of surfaces have been
created for modelling. This process had been completed by
selecting point clouds and creating fitting patches of them. If
the modelling of the big surfaces made from small patches,
because of the fitting problems some error can arise. The
patches have to be edited by user to solve this problem, and
then the patches can be fitted on their true surface. The 3D
editing possibility of software makes that easier. However this
process could be done with automatically mesh generation
process, but in this study the 3D modelling has been realised
only by using interactive method (Figure 5,6). Only the side
evaluated by photogrammetric method had been modelled.
Figure 6. Different perspective of 3D model
3.3 Comparing two methods
The 3D outlines created by photogrammetric system and 3D
modelling created by evaluating laser scanned data had been
overlapped (Figure 7).
Figure 7. Overlapped models
Coordinates of 60 specific points as corners had been measured
from two products. Residuals vx, vy, vz had been calculated by
difference of coordinates (Table 2).
ovx(m) ovy (m) | ovz(m)
0.0145 0.049 0.069
Table 2.Standart deviations of residuals
