International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
acquire, according to which kinds of topographic product must
be yielded.
Assumed as given the long-range TLS to be used (the
possibility of selecting between more than one instruments is a
chance that seldom happens!), 6 metrological parameters of
TLS may affect the result of the survey (see lavarone & Martin,
2003):
e horizontal and vertical FoVs;
e range measurement accuracy;
e horizontal and vertical scan resolutions;
e size of beam-spot in the range of involved distances.
We would like to proposed in next sub-paragraphs a workflow
for data acquisition strategy in case of a typical survey of a
geological site.
2.3.1 Preliminary site investigations: Disregarding all
concernes involved in the geological analysis of the site
(existing geological maps, evident indications and damages,
formulation of ground displacement hypotesis) and focusing on
the merely geometric survey, available large scale maps of the
interested area could be very helpful for the laser scanning
survey design. Possibly, maps should be integrated by
orthophotos and a DTM, in order to have a complete look over
the study area.
Collected information are needed to define the area to survey
and the area where TLS could be stationed. Furthermore, in
case the data acquisition has to be repeated in future for
monitoring purpose, some stable areas must be localized in
order to place some monuments to materialize a permanent
GRS.
2.3.2 Positioning of TLS stations: A survey design should
be first define positions of TLS stations, so that the whole
object coverage at requested spatial resolution and accuracy
could be guaranteed. To this aim it is necessary to compute the
ground resolution of each scan, given the TLS stand-point and
the position of the area to survey. Obviously, ground resolution
may change inside a same scan, depending on polar coordinates
(d, @, 0) of each measured points and on horizontal and vertical
scan resolution (Ag, A40). We introduce two relations which
allows to compute horizontal and vertical linear footprints (ry,
ry) as function of the above mentioned parameters. In order to
make a rough simplification of the real configuration, we
assumed a simplified terrain model consisting in a plane tilted
by a yangle, namely the vertical attitude used by geologysts.
Referring to Figure 1 for the meaning of symbols, horizontal
footprint ry can be computed by formula (1), where also
expression as function of the horizontal range d, is given;
during survey planning, the horizontal distance dy, can be
measured from maps instead of the slope distance d.
à Ke , d'Ap dip d,Ap
sin A2. a 2.0 7f. e (1)
3
cos Q 2 cosp 2 cosp sinOcose
Lf
According to Figure 1, the vertical footprint ry can be computed
as follows (both real and horizontal ranges have been
considered):
d ^0 d, A0
= (2)
: 2cos(y + 0) — 2sinOcos(y * 0)
n
Thank to the proposed formulas, for each scan position the
ground resolution can be computed and the requirements
needed for the information extraction verified.
HORIZONTAL PLANE
x
- :
-
-—
——
-
-—
-
-
-
-—
—
-
-
-—— ———"
- —
- -—
an
VERTICAL PLANE
-——
-——
-
-
_
—
-
-
-
-
-
——
-
-—
-—
———
——
-—-—
-———
-
- — —
—— —
4
M hp
TLS stand-point
Figure 1: geometric scheme for computing footprints.
Once the coverage of the ground has been checked, a rough
prevision of the amount of data to collect has to be done. One of
the main open problems in today’s laser scanning technique is
dealing with a very huge dimension of point-clouds, which get
difficult the data processing. Even though a resampling of each
scan could be performed after the acquisition on the field, the
best solution would be to directly carry out the data collection
at a suitable resolution. This strategy will result in a reduction
of the acquisition time as well.
2.3.3 Control point position and shape: According to
considerations made at par. 2.2, we consider the registration
method completely based on GCPs as the only which may
guarantee a sufficient accuracy in surveys for geological
investigations. This approach needs the positioning of at least 4-
5 control points for each planned scan. Even though the direct
topographic measurement of all CPs would be the most reliable
way (so that all points would play as GCPs), in practise this
approach might be largely time-wasting. Moreover, some CPs
could be needed in positions where their direct topographic
measurement by a total station of GPS would not be possible. In
case scans are linked to each other by TPs, these must be
positioned to avoid instability and bad determinability of
orientations. If the TLS allows an horizontal FoV of 360°, it is
possibile to put some GCPs also behind the stand-points,
perhaps in poisitions far from the landslide area where GPS
measurement can be carried out.
m If more acquisitions are planned at
different times for monitoring a landslide
evolution, some fixed GCPs should be
placed. We consider the permanent
positioning of a target as a not very
reliable solution for re-positioning of
GRS at each survey time. Indeed, GCPs
near the landslide area could be not
stable, while GCSs in external positions
could not be enough to give a strong
constraint. A suitable solution would be to
place some monuments in stable sites
outside the area directly involved in the
ground displacement. Starting from these
Figure2: ! T | us
reflective points, coordinates of required GC Ps
target for LR could be measured at each survey time by
TLS topographic methods.
applications Last but not least, the structure of targets
with GPS adopted as CPs is a critical aspect which
antenna should be accurately considered. Yet TLS
mounted producers usually proposed their own
targets (plane or 3D reflectors, such as
710
Inter.
cilinc
invol
topog
carrie
TPs,
mour
to fix
2.4 |
This
comp
the |
positi
cover
check
this,
resolu
modif
out di
Anotl
proce:
acqui:
Acqui
range
color
Digita
indep:
camer
corris;
image
carrie
et al.,
respec
In bot
by ma
and tl
calibre
points
25 D
Data p
In the |
derivec
registre
devote
we bri
the reg
geolog
