3.4 Results
3.4.1 Processing of the Image Time Series for Erbil
In this project, old sketches, historical (declassified) satellite data and new satellite
data were used. The aim of the project was to demonstrate the usefulness of
combining historical data for city growth documentation. The biggest developments of the city were recorded between 1950 and 1960 and after 2000.
Only a simple classification scheme similar to the one used for the Corine land
cover was used. With the use of satellite images, land uses are manually classified
into four classes: urban areas (red); parks (green); cemetery (yellow); and streets,
railways and airport (blue). The train station and airport are new dominants after
1980. For better classification, complementary data are needed. Although it is a
simple classification, it is a very well documentable enormous city growth and
outlines the directions of construction expansion. This can help in land use planning
and sustainable city development. The next pictures (Fig. 3.11) show the intensive
city growth of Erbil based on old plans or sketches and later on from satellite
images from the range of 1944–2012 is shown.
3.4.2 Creating a Base Map of Al-Qala Citadel
The second part of the project was to map Erbil’s historical centre (Al-Qala citadel).
The whole Citadel has never been documented with modern methods before. A
satellite image from the QuickBird satellite was used for the ground plan of the
Citadel (date of image acquisition: 2005-08-23). The satellite image was processed
by using Geomatica 10.0, ENVI Classic (pan-sharpening) and Adobe Photoshop
7.0. As mentioned above, the image sharpening, filtration and interpolating to
25 cm pixel were used for image quality improvements. The outputs of this
procedure are encouraging and enable the next step to process with better quality.
A provisory geodetic network in the Citadel area was built and over 600 object
points were geodetically measured. Next, 16 control points were signalized and
measured for aerial imaging, mainly on the roofs. The Photomodeler software was
used for all photogrammetric image processing (Pavelka and Bı ´la ´ 2013). Finally the
33 aerial images and 19 terrestrial images were adjusted to the base model. The
mathematically least square adjustment and absolute transformation utilizing the
control points were sufficiently accurate: the mean co-ordinate’s error of control
points was approximately 15 cm (mean position error 20 cm). In the 3D model
about 1,000 object points were measured and calculated. A comparison of
geodetically measured and Photomodeler calculated object points were made; a
small systematical and scale error had been found (the model from Photomodeler
was a little bit bigger and moved to south; the typical differences were approximately 15–30 cm in comparison to geodetic measurement by total station). These
56
K. Pavelka and E. Matous ˇkova ´
3.4.1 Processing of the Image Time Series for Erbil
In this project, old sketches, historical (declassified) satellite data and new satellite
data were used. The aim of the project was to demonstrate the usefulness of
combining historical data for city growth documentation. The biggest developments of the city were recorded between 1950 and 1960 and after 2000.
Only a simple classification scheme similar to the one used for the Corine land
cover was used. With the use of satellite images, land uses are manually classified
into four classes: urban areas (red); parks (green); cemetery (yellow); and streets,
railways and airport (blue). The train station and airport are new dominants after
1980. For better classification, complementary data are needed. Although it is a
simple classification, it is a very well documentable enormous city growth and
outlines the directions of construction expansion. This can help in land use planning
and sustainable city development. The next pictures (Fig. 3.11) show the intensive
city growth of Erbil based on old plans or sketches and later on from satellite
images from the range of 1944–2012 is shown.
3.4.2 Creating a Base Map of Al-Qala Citadel
The second part of the project was to map Erbil’s historical centre (Al-Qala citadel).
The whole Citadel has never been documented with modern methods before. A
satellite image from the QuickBird satellite was used for the ground plan of the
Citadel (date of image acquisition: 2005-08-23). The satellite image was processed
by using Geomatica 10.0, ENVI Classic (pan-sharpening) and Adobe Photoshop
7.0. As mentioned above, the image sharpening, filtration and interpolating to
25 cm pixel were used for image quality improvements. The outputs of this
procedure are encouraging and enable the next step to process with better quality.
A provisory geodetic network in the Citadel area was built and over 600 object
points were geodetically measured. Next, 16 control points were signalized and
measured for aerial imaging, mainly on the roofs. The Photomodeler software was
used for all photogrammetric image processing (Pavelka and Bı ´la ´ 2013). Finally the
33 aerial images and 19 terrestrial images were adjusted to the base model. The
mathematically least square adjustment and absolute transformation utilizing the
control points were sufficiently accurate: the mean co-ordinate’s error of control
points was approximately 15 cm (mean position error 20 cm). In the 3D model
about 1,000 object points were measured and calculated. A comparison of
geodetically measured and Photomodeler calculated object points were made; a
small systematical and scale error had been found (the model from Photomodeler
was a little bit bigger and moved to south; the typical differences were approximately 15–30 cm in comparison to geodetic measurement by total station). These
56
K. Pavelka and E. Matous ˇkova ´