46 
Figure 3 - GPS frame network 
Dimensions of error ellipses obtained by adjustment, at a 
confidence level of 95%, have a maximum value of 11.7 
cm for major semi-axis while the confidence interval in 
height at the same level is variable from 2.6 cm to 18.7 
cm. 
Five points which are known in both systems were used 
in order to transform the co-ordinates from WGS84 
system to Italian one; the points are: three 3D points, one 
2D points, and one ID point (the benchmark). 
A Molodensky transformation was used. Eight 
“planimetrie” equations (in latitude and longitude) and 
four “altimetric” ones (in height). 
Transformation was carry out: 
without using a model of geoidal undulation (by 
using orthometric heights as “local” heights) 
by using the local undulation model called 
ITALGE095, by Polytechnic of Milan (by using 
the ellipsoidal heights as “local” heights) 
Since the zone is very small, the differences between the 
calculus results are imperceptible. 
Then a software was wrote with the aim to estimate the 
variance components of the co-ordinates used in 
Molodensky formulas. 
Observations are separated in two groups: the planimetrie 
(latitude, longitude) lj and the altimetric ones (height) 1 2 . 
The linear partitioned model is: 
A" 
V 
v, 
A. 
X = 
_-2_ 
+ 
_—2_ 
with the variance matrix 
X = a 0 2 Q = o, 2 V, + o, 2 V 2 
The variance relative to each group of observations was 
estimate: 
s 1 2 = v i ‘Py I /tr(WV i ) 
In order to verify the convergence, the calculus was 
repeated with different approximate a-priori values of 
variance: 
a o = a * = 0A m 
a h ranges from 0.001 m to 5 m 
Estimated standard deviation has been: 
s 0 = s^ = 3.0 cm ; s h = 3.7 cm 
The resultant residuals of the transformation are very 
low; they indicate the strong congruence of the points and 
the goodness of the network even if the transformation 
redundancy is low. The estimated parameters of the 
transformation will be used in order to transport into the 
National cartographic system the co-ordinates of all those 
points which were measured by means of GPS 
technology and connected to the frame network. 
4. GR< 
Ground 
cartograp 
In order 
landslide 
five year 
This map 
Cassa d 
abbreviai 
the photc 
The area 
heights v 
and the 
height ab 
The cor 
digitisatii 
So it's be 
- The 
throi 
to th 
A se 
lines 
softv 
inter 
and 5 
- The 
trans 
For one r 
more rec< 
municipa 
Unfortun; 
had not b 
Cartograp 
contour 1; 
derive a 1 
GRID wa 
points we 
5. COMÍ 
Once the 
landslides 
and estim 
Unfortum 
from the 
absence ii 
The lines 
pass at a c 
Problems 
the p 
made. So 
enough fc 
case of us 
with dista 
level at a 
where the 
placed at¡ 
the re