XXIX-B8, 2012
3) according Spanish
terrain instability
ind man-made slopes
movement in which
ough with a certain
carp can be observed
pening). The flow of
er can be observed in
cut), but the typical
is not developed by
the very presence of
he marly materials,
pond to a geological
of Lower Cretaceous
ic of External Zones
t background in the
sic Age, underlaying
iths, practically until
ation of Civil Works
npaigns were carried
weeks period, in the
d landslide affecting
side the red square.
NER SURVEY
e mass, two scanning
LS in a time interval
tech ILRIS 3D long
eometry of the scans
ig the whole area and
fa transference to the
in the first campaign
campaign have been
ocation of the scan
the correct geometry
system (points well
» straight line), the
accessibility.
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B8, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
Besides, the presence of overlapped zones in the scans must be
taken in account because the adjustment of the different point
clouds is based on the use of surface matching. The point
resolution in each capture was 5-6 cm at the average scanning
distance, that allows obtain a point cloud of higher density once
the points clouds are merged.
Figure 3. Position of the scan stations and final result of the
point clouds merging in the two campaigns.
3.2 GPS measurements
To georeferencing the scan stations, a GPS receiver (Leica
System 1200) is placed on the laser scanner instrument by
ILRIS GPS mount kit set. We can relate both systems and the
data captured by them because the distance between antenna
phase centre and scanner origin is known (figure 4).
Figure 4. GPS receiver with the ILRIS GPS mount kit set
placed on the scanner. Scanner reference system
and offsets (dy, dz) between antenna phase center
and scanner origin are represented.
23
4. DATA PROCESSING
41 GPS data
GPS observations were processed using GeoOffice software
(Leica Geosystems). Every antenna phase center is calculated in
the ETRS89 reference system with a sub-metric accuracy.
4.2 Point clouds (TLS data)
First, the position of antenna phase center was incorporated to
the point clouds as an additional point in the TLS reference
system, taking in account the offset of this point to the TLS
instrument origin (figure 4).
Next, the relative orientation of the different point clouds of
each scan station is made, adjusting and merging them in a
single point cloud using the module “matching surface features”
of I-Site MAPTEK software. This orientation is based on the
correlation of some common zones in the overlapped point
clouds (figure 5). These common areas are recommendable to
be cleared to avoid noise and outliers which can affect the
matching.
In this way, we obtain for each campaign a single point cloud in
a local reference system. The average density of these point
clouds is about 1 point every 2 to 4 cm in zones without
shadows, corresponding to 5.800.000 and 8.600.000 for the first
and the second campaign, respectively.
4.3 Surface models and DEMs
After single point clouds corresponding to both campaigns are
obtained, a surface model (DSM) can be obtained where we
take in account both the points assigned to terrain surface
(terrain points) and the points not assigned to the terrain surface
(non-terrain points). These last points correspond to vegetation,
fences, constructions, vehicles and other features located over
the terrain surface or noise points introduced by in motion
vehicles, birds, rainfall drops or other issues (figure 6).
Then we proceed to point classification or filtering in two
classes, terrain and non-terrain points (Lichun, et al., 2009).
From the classified point clouds in which only terrain points are
present and non-terrain points are deleted, a TIN model of
terrain surface or DEM is obtained.
Figure 5. Relative orientation of single points cloud using
common or overlapped zones.
