×

You are using an outdated browser that does not fully support the intranda viewer.
As a result, some pages may not be displayed correctly.

We recommend you use one of the following browsers:

Full text

Title
Proceedings, XXth congress


International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B-YF. Istanbul 2004

11 bits per pixel that correspond in 2!! = 2048 layers per pixel
contrary to the 2° = 256 layers that provide the traditional
images with 8 bits.
e A surveying diagram of the same area in scale 1:5000
of the year1978of suburban region of municipality Pylaia in
scale 1:5000. The region was selected to be suburban for
facilitation reasons with few buildings and soft slope for a better
accuracy achievement. For the digitization of control points the
above-mentioned map was used.
e The Digital Terrain Model (DTM), of which the
digitalization step and accuracy were one meter (Ntokou K. -
Theodorou M., 2001).
e Finally, the coordinates of 7 points, in Hatt projection, of
the national network were used, in order to transform the
measurement from the GPS, which were in the projection
system WGS'84
3. IMAGE ORTHO RECTIFICATION
3.1. Image ortho rectification with control points digitized
from surveying diagram
3.1.1. Rectification
For the image rectification 22 points were used (13 control
points and 9 check points) the coordinates of which were
received via digitization of map in scale 1:5000, which it was in
Hatt projection. The polynomial model was selected and more
concretely Afine transformation. (Erdas, 2001)
The RMS error for the control points was 0.4429 m, which was
analyzed at X equal with 0.3508m and at Y equal with 0.2704m.
Similarly the RMS error for the check points was 0.0920m,
which was analyzed at X equal with 0.0713m and at Y equal
with 0.0582m.
Considered these errors bearable, the next step was the image
resampling with the Cubic Convolution method that resulted to
a rectified image (Tsakiri —Strati M., 1998).
3.1.2. Ortho rectification
The image rectification was followed by ortho rectification with
the use of Digital Terrain Model (DTM) of the region, the
control points and the choice of model of satellite Spot. (Mpanta
K., Xalkidou M., Xalkidis L., 2001)
22 points were used (13 control points and 9 check points). The
RMS error for the control points was 0.4114m, which was
analyzed at X equal with 0.2929m and at Y equal with 0.2888m.
Similarly RMS error for the check points was 0.0549m that was
analyzed at X equal with 0.0264m and at Y equal with 0.048 1m.
Finally the total vertical error of rectified image's adaptation in
the DTM was estimated equal with 0.356 m.
With those errors considered bearable, the final step was the
image resampling with the Cubic Convolution method of the
rectified image and the production of the orthoimage (Figure 2).
3.2. Image ortho rectification with control points measured
with GPS
3.2.1. Rectification
For the image rectification 20 points were used (12 control
points and 8 check points) that were received from
164
measurements, which were realized with the Global Positioning
System (GPS).
The RMS error for the control points was 0.3685m, which was
analyzed at X equal with 0.1822m and at Y equal with 0.3203m.
Similarly the RMS error for the check points was 0.0920m,
which was analyzed at X equal with 0.0497m and at Y equal
with 0.0774m.
Considered these errors bearable, the next step was the image
resampling with the Cubic Convolution method and the result
was a rectified image.
3.2.2. Ortho rectification
The image rectification was followed by the ortho rectification
with the use of digital terrain model (DTM) of the region, the
control points and the choice of model of satellite Spot.
20 points were used (12 control points and 8 check points). The
RMS error for the control points was 0.3243m, which was
analyzed at X equal with 0.1640m and at Y equal with 0.2797m.
Similarly RMS error for the check points was 0.0962m that was
analyzed at X equal with 0.0810m and at Y equal with 0.0519m.
Finally, the total vertical error of rectified image's adaptation in
the DTM was estimated equal with 0.507 m.
With those errors considered bearable, the final step was the
image resampling with the Cubic Convolution method of the
rectified image and the production of the orthoimage. (Figure 3)


Figure 2. The panchromatic orthoimage using control
points digitized from map.