International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004 
  
  
  
  
  
  
  
  
  
  
  
| Ini 
| — 
2000, KVR-1000 images for 1988 and airborne photographs for software, coordinate transformations by an Excel calculation, | 
1968 and 1940’ies. The first time period is mentioned as The first georeferencing of maps is a normal Affine- 
1940' ies, because the airborne photos for whole study area had transformation to place the scanned maps to the position of the 
not been taken in the same year. They had been taken in maps origin coordinate system to enable a good orientation and TI 
DD . . 95 . x . ] 
different periods of 1940'ies between 1942 and 1949. to detect more easily the displacement in the next step. That 
- . + . ge 
was only the first and approximate transformation and done by dil 
: : NE. di 
2.2 Study Area using the map corners. Afterwards, all the maps are transformed | 
: ro 
exactly to the corresponding coordinate system using all the 
Study area has been selected not only as the city center ol grid points on them. With this partial transformation, the maps 
Istanbul, but also containing the surrounding territories of are not only georeferenced also the errors are reduced. With this 3.1 
Istanbul, which can affect the landuse and its change as well as transformation, perfect neighborhoods and a mistake Th 
their interpretation and analysis. The study area contains two minimization of scanning and paper sheets (and sometimes also 
zones, the center and buffer zones, as illustrated in Figure 1. of the print) effects have been got. Then a third Affine- 
transformation, to move the image to the other coordinate 
system (from one to another strip or ellipsoid), had to take 
place. The correct coordinates have been calculated by the map- 
corner coordinates or by the grid on the map-sheets. 
After georeferencing the topographical maps, satellite images In 
were georeferenced using image to map method. In that way, all act 
the satellite images were ready to digitizing process. Following do 
process was the creation of orthophotos for 1968 and 1940'ies. 
Totally 94 orthophotos for 1968 and 114 orthophotos for 
1940'ies were produced. 
2.4 Map Production 
Manual method was selected because the richness of legend, in 
which 62 different landuse types was defined in total four 
— levels. This caused a very long and hard work in map 
Figure 1. Center (grey) and buffer zone (light grey) with water production. Each of four landuse detection step has three layers. 
surfaces (dark grey) The first one is deals with linear objects such as transportation 
2 facilities and rivers where the second one deals with areas and Wi 
2.3 Data Preparation {he third with 3D areas such as bridges, viaducts or tunnels. As ; 
; : : : an overview, Figure 2 shows the manual digitizing and X 
Data preparation has been begun with the year 2000, which will aera ite du : E = 
: : = ^ classification results for the first level of legend. 
be mentioned as reference year. Map to image method was Wi 
selected for georeferencing the satellite image. All maps have dec 
been scanned with 600x600 dpi geometric and 24-bit 
radiometric resolution using a HP 4C/T DIN A4 flatbed Th 
> - . : - ( 
scanner. Georeferencing has been done by using TopoL-GIS 
Viwier Ehurlaces e. 
Age Anas 
Asifsétal oslsdun c 
3 €, 
| Tessa arel 
Burr idu Aredà 
c, 
wh 
dire 
vali 
Cir 
CP, 
Me: 
MS 
Circ 
RnR : mana : ; ; At … 1968: CM 
Figure 2. Results of the digitalization and manual classification for all four evaluation years. (left upper: 1940’ies; right upper: ; 
left lower: 1988; right lower: 2000) 
  
1026