International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
As a result of the wavelength of the LIDAR laser signal of 0.9
um, the intensity image is an infrared image.
[t was considered that the evaluation of the planimetric
precision of the intensity image is important for two reasons:
- to evaluate its one spatial precision;
- to evaluate the planimetric precision of the LIDAR data.
A set of points representing well defined points on the intensity
image and on an ortho-photo used as reference. The ‘reference’
ortho-photo was produced from the digital images recorded by
the Mosaic Mapping LIDAR system using a set of GPS check-
points as control-points and the LIDAR generated DEM using
the Leica Helava DPW. Its ground resolution is 25 x 25 cm. The
results of this evaluation are shown in Table 2. [t seems that the
planimetric precision of the intensity image is determinate
mainly by its resolution of 1x1 m.
Std. dev. (m)
X 0.50
y 0.65
Table 2: Planimetric precision of the intensity image
4.3.4 Rapid production of orthoimages
Taking into account the needs of emergency mapping users, an
evaluation was done about the rapid production of geo-spatial
data from the on-board digital camera with out using ground
control points. The goal is to use only the data produced by the
LIDAR system and its coupled digital camera. A test was done
using 6 digital images from two adjacent flight lines and the
‘direct orientation’ (DO) values computed from the GPS +
inertial sensors of the LIDAR system and the LIDAR produced
DSM. Two possibilities were explored, the stereo-compilation
and the production of ortho-rectified images. For this reference,
the same digital photos were used following the standard
method (relative and exterior orientation) using the Leica
Helava working station and the GPS recorded control points.
a.- Stereo-compilation: The stereo compilation was not
possible because the residual y parallax was much to important.
Comparing the values of the photogrammetric computed values
of the exterior orientation angles with the ‘direct orientation’
values it appears that these last values have errors of some
minute of arc. The differences for the Omega and Phi angles
were in a range of 1 minute to 16 minutes of arc with a mean
value of 8.25 minutes. For Kappa the differences were smaller,
between 2 and 4 minutes.
b.- Ortho-rectification: A digital ortho-photo mosaic was
produced using the same data set with the ‘direct orientation’
values. As expected, planimetric errors in order of some meters
were observed. The relative errors, measured between the
position of the same terrain detail present on two successive
photos were in the range of 2 m to 4 m. These errors are easy to
observe on the ortho-image at the seam line, where the image of
a linear detail as a road will appear as ‘cut’. On the example
shown in Figure 2, the road on the left of the image appears to
be cut in the middle indicating a relative orientation error
between the two images. The red lines are the GPS measured
road centre-lines and their offset to the East of the ortho-image
shows an absolute error in the East-West direction. The
absolute error, measured between a detail position on the ortho-
image and the GPS position measured for several points had
values up to 6 m.
It is understood that for this ‘first shot’, automatically produced
ortho-image, the absolute planimetric precision could only be
estimated from the ‘a priori’ precision estimation of the exterior
orientation parameters delivered by the GPS / inertial sensors of
the LIDAR system.
These ortho-images can be very useful for emergency mapping
operations despite the much higher errors versus the normal
acceptable values for this category of image scale. In fact, for
the firs step of situation understanding and preliminary damage
evaluation after en emergency event, the value of a new, quick
delivered, very high-resolution colour ortho image is very high
despite an absolute positioning error of about 2 to 6 meters.
In second iteration, the orientation operations can be done again
using additional information and more precise ortho-image can
be produced in a few hours.
5. RECOMMANDATIONS AND CONCLUSIONS
5.1 Recommendation for the use of LIDAR data and rapid
change detection for emergency mapping
While in the imagery the unit is the 2D pixel in image space, in
the LIDAR data the unit is a 3D point in the space of the
reference coordinate system. The LIDAR data is a high-density
point sampling of the terrain. These data are processed either as
TIN elevation surface model or interpolated into a regular
elevation surface model (DSM). The LIDAR DEM data can be
also converted into a raster image. In addition certain LIDAR
systems capture also the intensity of the response signal
corresponding to the LIDAR ground point. The intensity is à
function of the reflectivity of the ground material and the
intensity changes form a georeferenced grey level image like
output.
In emergency situations, where the time factor is crucial,
LIDAR data can be used for mapping, modelling, change
detection and monitoring and visualization tools for knowledge-
based decision-making.
In mapping, LIDAR data contributes to an accurate recording of
the elevations (shape/morphology) of the terrain surface, the
generation of a DEM (bald earth, contours) by separating man-
made objects and vegetation from the ground surface, the
calculation of terrain information (slope, aspect, volumes,
profiling), the ortho-rectification of imagery, the extraction of
planimetric features (i.e., buildings, Figure 3 and 4) in fusion
with other data sets, and the use of the georeferenced intensity
image for mapping (c.g., roads) and the detection of high
reflective objects.
Intern
deri
Figur
In moc
models
during
3D vi
unders
assessi
draped
If prey
acquire
for ch:
972
— EEE E emunt te t m
