133
Figure 2 - Planimetric diagram of the topographic survey
of the 1998 network
Point
Vx (mm)
Vy (mm)
Vz (mm)
1
3
2
0
2
4
2
-2
3
0
3
-2
4
-1
-4
0
5
3
1
-1
6
-4
1
-5
7
0
5
-2
8
2
-2
2
9
-2
-5
4
10
2
0
-2
11
-5
-2
5
12
-1
-3
-3
13
0
1
3
14
-3
2
4
RMS
3
3
3
RMS Vxyz = 2.8 (mm)
Table 1 - Residuals of the comparison
Profiles and DSM comparisons
Horizontal and vertical profiles of 1992 and 1998 have
been realized in the same planimetric positions of the
cloister facade, through the parameters of the affine
transformation calculated before (table 1).
The horizontal profiles are shown in the figure n. 4.
OO, 02, 03 are the mains horizontal profiles realized at
different heights. From the bottom of the facade, we have
OO; 02 is the median and the 03 is the upper one (figure
3). It is evident that the profiles 02 and 03 are placed one
upon another, while the horizontal profiles 03, have
significant differences starting from the 111 meters of x-
axis. This trend is confirmed in the following comparisons
and that it is imputable to the placing of the tie beams at
first floor down side the roof.
Generally the differences between profiles concern the
two surveys are of about one centimeter in a zone around
the zero. Looking the diagram concerning the section 02,
the differences are distributed around the zero. While the
differences concerning the horizontal profiles 03 we can
strangely see that the trend of the points is a curved line
(figure 5).
The value is negative when 1992 is greater than 1998.
We can remark a moving-back of the top part of the
facade, increasing towards the center.
The vertical profiles of the facade are compared in a
diagram where the depth is in the abscissa and the
vertical axe parallel to the front is in the ordinate (figure 6-
a).
The trend of the four profiles is similar. Looking the profile
sample V2, the most important shiftings are in the upper
part of the cloister facade, that is the loft. The punctual
differences of the vertical profiles are in a range of about
one-centimeter around zero (figure 6-b). The most
important differences are negatives and show a moving
back of the upper part of the facade. The shiftings are
similars even for the other vertical profiles.
A DSM with 30 centimeters squared grid was realized with
the analytic DIGICART 40 stereoplotter, to have a spatial
vision of the deformations of the cloister facade.
Figure n. 7 shows the punctual differences of DSMs
realized with 1998 and 1992 surveys.
Negative punctual values show a moving back of the
facade. The most important differences are noticed in the
loft and they have a maximum value of 5 centimeters.
DSMs raster of sampled areas
To obtain a visual comparison of the restoration
performed, two orthophoto of masonry details concern
1992 and 1998 photogrammetric surveys has been
realized.
Glass plates were scanned with a 1000 dpi resolution by a
photogrammetric scanner. Classic method for photograms
orientation has been adopted. It consists of three phases:
the inner orientation, relative and absolute. The residues
of the absolute orientation have been of 3-4 millimeters in
three coordinates.
With regard to the orthophoto, first of all, we have carried
out the restitution of the area concerned (figure 8).
The break line so inserted and some tridimensional points
placed manually, have been the bases for the automatic
generation of a DSM with regular 10-centimeter grids. The
Stereo Matching algorithm using an appropriate
correlation coefficient do not plot all the point of the grid,
but only someone. Than the data of this DSM were
interpolated to enable the orthophotos to be constructed.
The orthophotos of the meshes, has been obtained. In the
figure n. 9-a concern a masonry detail of the 1992 survey,
the restitution has been put on the orthophoto. Remark a
vertical crack in the right hand side of the photo.
In the figure n. 10-b concern the same building part, after
the repair work. Even if the photo is more exposed, we
can see that the bricks have been cleaned by the mortar
and the crack has been closed.
