International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
Italy. The interest of this survey is in the fact the disaster has 
not yet happened, and the acquired data will be used to design 
works for consolidating and preventing a possible landslide. 
After a long period of intense rain at the end of 2002, water 
infiltrations has begun to corroded the bottom sector of the 
slope, calling for interventions of surveying and protection. 
The studied slope is about 70 m in length, 45 m in width and 50 
m in height. 
In order to define a permanent GRS for multi-temporal 
acquisitions, 4 monuments have been placed on the other side 
of the small valley. Relative positions of these points have been 
determined with a sub-cm accuracy either by static GPS 
measurement and by a polygonal network by total station. 
Thank to the avilability of two GPS permanent station of the 
Politecnico di Milano located in the research centres of Como 
and Lecco, coordinates of monuments have been also referred 
to the national geodetic network IGM95, so that the whole laser 
scanning survey can be inserted in existing maps. Thank to the 
positioning of these points, we have then placed targets in 
positions to assure a good stability for 3D-views registration, 
disregarding the fact they could be in sites subject to movement 
in the future. In fact, 8 targets have been placed, 5 of these 
consisting in a special structure (Figure 2) mounted on a tripod, 
made up of a disk covered by a highly reflective material and 
by a plug to fix a prism of a GPS antenna; two different 
diameters for targets are possible on two side of the rotating 
plane (20 and 50 cm). Positions of these targets in the GRS 
have been determined by intersection from monuments, so that 
they can be considered as GCPs. Furthermore, 3 circular small 
reflecting targets (diameter of 8 cm) have been placed in the 
central part of the surveyed area to register a detail scan to the 
other ones; they are not known in the GRS, playing the role of 
TPs. 
A Riegl LSM-Z420i laser scanner has been used (more 
information can be found at website of Riegl). Three scans have 
been acquired, two (scan 1 and 2) giving a general view of both 
sides of the whole slope, the third (scan 3) focusing on a detail 
of the critical central part where a landslide could start. 
All acquired 3D-views have been processed using the RiSCAN 
PRO Riegl SW, used also for scanner control. Processing has 
involved a pre-editing of each scan, consisting in removing 
trees and marginal areas, in resampling to reduce the number of 
points, as well as in automatic measurements of reflective 
targets. Registration has been carried out by using 5 GCPs for 
both scans 1 and 2; then scan 3, where only three GCP could be 
view, has been pairwise registered to scan | by introducing also 
TPs. A mean sigma nought of 2 cm for each registration 
(directly to GRS or pairwise to another scan) has been reached. 
After each scanning, some images have been collected by a 
Canon EOS 1Ds (11.1 Mpixel) digital camera, fixed on the top 
of LSM-Z420i TLS in stable way. Once the camera has been 
positioned in a such manner, the RiSCAN PRO SW allows to 
register images to 3D-view. By using an approximate 
registration, the user can view the position of reflective targets 
(or some other features manually extracted) projected into the 
images; by simply interactively correcting residuals between 
position of targets on the image and their projection from 3D 
scan, the registration parameter from image to scan can be 
computed. However, the intrinsic calibration of digital camera 
must be known. The advantage of using a camera already 
registered to the scanner results in the possibility of colouring 
the 3D-models or obtaining a photo-texture representation 
without any further task; in Figure 3 a colour 3D model of the 
whole study site is reported. 
  
  
  
  
  
Figure 3: frontal view of the landslide in Caslino d'Erba; the 
point-cloud have been added up by the colour information. 
3.2 Botticino Marble Quarry (Brescia) 
A Marble Quarry in the Botticino area (Brescia, Italy) has been 
selected to test the TLS technology for geological survey 
(Monti et al. ,1999). Main objectives of this test have been: 
e defining the technical procedure for the usage of TLS 
in quarries; 
e testing laser scanning approach for geological 
applications. 
Field survey campaign is organized in the following main steps: 
1 definition of TLS stand-points; 
2. target (GCPs) displacement; 
3. acquisition of range scans and digital imagery; 
4. georeference of targets by GPS. 
Two TLS stand-points have been selected outside the quarry so 
that both excavation area and GCPs were well visible, and in 
the way that extraction activity would not be interrupted. (Fig. 
4). 
Scans obtained from different viewpoints have been 
transformed into a single reference frame (Fig. 4) using highly 
retro-reflective targets built for LRTLS applications (Fig. 2). A 
GSP antenna mounted over the target has been used for 
georeference. The shift between antenna’s phase centre and 
retro-reflective target centre has been previously measured. 
During laser data acquisition, digital image of the quarry have 
been acquired as well. 
  
  
Figure 4: Range scans obtained from different viewpoints (1,2) 
are transformed into a single reference frame. 
712 
  
Int 
Da 
car 
ma 
La: 
Re 
Co 
foll 
Pre 
incl 
a ve 
Af 
soft 
whi 
Star 
refe 
be | 
pro 
com 
life. 
An 
reso 
3D 
soft 
colo 
geol 
The 
quar 
infoi 
whic 
and 
repre 
Mak 
been 
meas 
creat 
carto 
This 
for g 
area, 
  
Figt