CERING
E
)0 Florianopolis,
establishing as a
iickly dense and
as in projects of
he time of post-
rst surveying for
a corridor with
ophotos with the
à obtained in that
chosen a stretch
rtophotos mosaic
1e environmental
decades brought
s Airborne Laser
Detection and
rnative for DTM
R make possible
:imeters, the that
projects. when
ds. LIDAR also
Digital Surface
ents technology
)viários, besides
ject, is still small
DAR technology
Thus, this study
ogy, and b) the
LIDAR products,
for the thematic
nately 1700m of
Luis Alves-SC-
ghway will have
nite sections in
igar the paved
uth of projected
of study area.
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
Figure 1 — Orthophotos mosaic of study area.
3 METHODOLOGY
Thus, this study aim at to contribute for a) LIDAR technology,
and b) the evaluation of your utilization. For this was made use
of LIDAR products, of orthophotos mosaic and of both
integration for the thematic analysis of a highway stretch.
Research available materials were: pos-processed LIDAR
products file, flight altitude 1,000m, realized in November of
2002; digitasl ortophoto mosaic, flight scale 1:15,000 and
scanerized with 0,40m pixel size realized in March of 2002;
aerophotogrammetric retitution in 1:5,000 scale.
Steps developed during the research they are described forward.
Therefore after, each one is explained.
3.1 LIDAR data treatment;
3.2 land use surveying using photointerpretation;
3.3 generation of Digital Terrain Model (DTM) and Digital
Surface Model (DSM);
3.4 creation of slope map;
3.5 thematic analysis of study area.
3.1 LIDAR data treatment
LIDAR data are in three groups: a) calibration data and
installation parameters (obtained before the flight), b) measures
of laser distances with your respective scanning angles and c)
POS data. Those data are processed and integrated, being
obtained at the end of this stage a LIDAR point cloud,
traditionally presented three-dimensional coordinates in WGS-
84 system and LIDAR pulse registration. To differentiate which
information correspond to relief or any other geographical
phenomenon or object present in studied surface, is necessary
to accomplish a data treatment. In case of this study, treatment
was accomplished in three main stages: filtering, classification
and manual edition of points cloud.
The filtering and classification definition happens in agreement
with objective to be reached and not with employed method.
Removal of underisable points is considered a filtering
operation. Already the task of finding a specific geometric or
statistics structure, as buildings or vegetation, is defined as
classification (AXELSSON, 1999). Steep of manual edition was
added to data treatment being taken in consideration that the
automatic filtering and classification method for own algorithms
for this end were not capable to present satisfactory results.
To generate a DTM that represents the bare earth in the closest
way of the reality, the correct definition of this surface in
LIDAR points cloud is indispensable. This does of treatment a
very important task and that influences excessively in DTM.
final quality.
Figure 2 show the developed flow chart for LIDAR data
treatment and cach stage explanation in sequence.
TREATMENT OF LIDAR DATA
1) POS-PROCESSING
LIDAR POINTS CLOUD
Y
2) STUDY AREA
CUTTING
3) FILTERING
4) CLASSIFICATION
i
5) MANUAL EDITION
pe
BARE EARTH LAYERS
BUILDING
Y
VEGETATION
OTHERS
Figure 2. Flow chart of LIDAR data treatment.
Treatment of LIDAR data:
1) Pos-processed LIDAR point clouds: all LIDAR datas Pos-
processed are considered properly georreferenciados without
filtering or additional analysis.
2) Study area cutting: initially, LIDAR points files were cut out
to coincide with orthophotos mosaic that defines direct
influence area of environmental impacts. The cutting resulted in
a file with 582.407 points.
3) Filtering: an automatic filtragem LIDAR points cloud was
accomplished aiming at separating bare earth points and object
points. Filtering was made in TerraScan program, that possesses
a specific tool for this task. The parameters for tool use (terrain
angle, interaction angle and distance) were defined based in
result analysis of dozens filtering tests accomplished in study
area being used different parameters.
4) Classification: in LIDAR points cloud classification defined
in filtering as not belonging to bare earth three layers were
created separating principal elements found in study area:
vegetation, buildings and others (transmission lines and towers).
The main objective of classification went to aid to find points
belonging to bare earth that were erroneously defined as objects
in the filtering process.
5) Manual edition: in this stage, group of automatically filtered
and classified points were analyzed in ArcView GIS. For
AXELSSON (1999), in many cases are impossible to LIDAR
data interpret unless images are available. The laser points were
put upon to LIDAR intensity image and orthophotos mosaic to
identify possible erroneously filtered and classified points for