The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008
2.2 TLS measurements
Terrestrial laser scanner measurements were perforrmed during
the GB-InSAR monitoring to obtain a precise and updated
DEM of the landslide.
A Riegl LMS-Z420 laser scanner, which allows long range
measurements (up to 700 m) and centimetric precision in the
2.5 cm range, was employed (Fig. 8).
The laser scanner was installed a few metres from the GB-SAR
thus ensuring good visibility of the landslide and of the markers
installed in the surrounding area. These markers are necessary
to georeference the acquired data and their position was
acquired through topographic measurements (see section 2.3).
The laser scanner data acquisition lasted about 40 minutes for
each of the two scans necessary to describe the landslide and
about 33 million points were acquired.
The global point cloud used to generate the 3D model of the
landslide was obtained after filtering and registration of the two
scans.
Figure 8. The Riegl LMS-Z420 Laser scanner
The laser scanner data were processed with a new Sirlo
software realized thanks to a collaboration between the SIR
society (Italy) and the Geomatics Research Group of the
Politecnico di Torino.
The laser scanner was firmly connected to a calibrated digital
camera (the Nikon D1X camera). Many digital images of the
scenario were acquired and they were employed to integrate the
laser scanner data with their radiometric content.
Thanks to the Sirlo software it was also possible to generate
solid images of the landslide (Bomaz et al., 2003), which allow
direct measurements to be easily made on the point cloud, such
as distances, angles, plane orientations,..
A portion of the processed laser scanner data is reported in
Figure 9.
Figure 9. Portion of the processed laser scanner data
2.3 Topographic measurements
In order to correctly locate the radar and laser scanner data it
was necessary to create a local coordinate system made up of
three control points and many markers. The coordinates of the
control points were estimated using relative-static GPS
positioning, while the coordinates of the other points were
acquired by the total station. This latter instrument was also
employed to measure the positions assumed by the active radar
calibrator that are necessary for the radar data georeferencing.
All the topographic measurements were adjusted with a
commercial software, and centimetric accuracy was obtained in
the estimation of the coordinates of the measured points.
2.4 Results of the campaign
The final SAR interferometry product consists of displacement
maps of the scenario, which provide information about the
distribution and entity of the movements.
The interferometric maps were obtained through a masking
procedure that excludes areas with lower coherence than the
reference lower limit value (0.9 in this case) and by focusing the
radar images on a DEM of the scenario.
Three displacement maps - corresponding to the three
sequences of acquired radar images - were obtained. One of
them is reported in Figure 10.
Figure 10. Displacement map relative to the first sequence of
SAR images
No appreciable displacements are present in any of the maps as
could be foreseen; the landslide is in fact quiescent but at risk of
reactivation and no precipitations occurred during the
monitoring campaign.
It will be possible to repeat the SAR measurement campaign
four-six months after the first campaign in order to determine
any displacements that have occurred during the time interval
between the two campaigns. In this case, it will be necessary to
adopt the “permanent scatterers” (or “coherent points”)
technique (Ferretti et al., 2001).
This technique is based on the selection of stable coherent radar
image pixels - between successive campaigns - whose
interferometric phases can be considered only affected by
atmospheric variability and any spatial movements.
The final products of laser scanner measurements are 3D
models and solid images of the scenario. In this case, we
obtained a complete 3D model and four solid images of the
landslide. Only on the upper part of the scenario laser data was
the density not so good because of the long distances, which
364