The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
1040 
containing description of the basic reference unit for 
digitalization, photo interpretation and digitalization rules. Each 
reference parcel has to have a unique code in the frame of whole 
Turkey which can be automatically attributed by the GIS 
software. Due to frequent changes of administrative ordering 
within Turkey, unique code is derived from geographical 
coordinates. 
Next step was to test the usefulness and effectiveness of ortho 
photos and the LPIS database to locate and define the 
boundaries of the current year’s agricultural parcels used by the 
farmers through consultation with the farmers on both pilot 
areas. The main objective of the consultation phase was the 
contact with the selected farmers and more specific to facilitate 
the location of their used parcels on the orthoimages, register 
them on LPIS database and to ask farmers to answer to a 
submitted questionnaire. Some interesting statistical figures are 
prepared as result of the consultation campaign, concerning 
differences in the areas declared by the farmers and those 
actually measured on the orthoimages. Following the 
consultation with the farmers in the field the outputs of this 
work was used in the database updating process. 
Within the frame of the Project, a specific SW Turkish IACS 
Demo Software (TIDS) was created for the demonstration and 
simulation of selected IACS procedures. TIDS was one of the 
basic outcomes of the project aimed to implement project data 
for producing real results for IACS testing procedures in the 
Tekirdag area and to train MARA staff for global IACS 
procedures. Real geographical data, real data of farmers’ 
declarations and real on-the-spot control data are all integrated 
within the SW. Any simulation allowed because the objective 
was to produce real statistics about eligible land area and 
declaration anomalies. 
3. EXPLANATION OF METHODS 
3.1. Imagery acquisition and orthophoto creation 
Very high resolution (VHR) Ikonos satellite images with a 
resolution (pixel) lm for the pilot areas were acquired from and 
provided by INTA Spaceturk in the vast majority between 
spring and autumn 2006. 
For the creation of the orthophotos the Erdas Imagine software 
was used. Every frame of acquired image was orthorectified 
using the Ikonos model and the rational polynomial coefficients 
(RPC) of the image. The precision of the RPCs was improved 
by applying a second order polynomial adjustment using precise 
ground control points (GCP). Sharp comers clearly visible in 
the image and well identified in the field (mainly cross roads) 
were selected to be measured in the field. In order to achieve a 
homogeneous distribution of GCPs within each pilot zone, a 
grid of 14 Km x 14 Km was used and one (1) GCP selected in 
every cell of the grid. The selected GCPs were then measured 
with geodetic GPS instruments to have a horizontal accuracy of 
about 20 cm. 
The digital vector elevation maps of 1/25.000 scale topomaps 
were used to create the digital terrain model. The “create 
surface” tool of Erdas Imagine was used to create the digital 
terrain model (DTM) with 5m grid size. 
The Ikonos level 2 scenes, with an average size of 300 km 2 
were orthorectified using an average number of 20 GCPs per 
scene. The average root mean square error xy (RMSE) value of 
the orthorectification procedure obtained was <1,70 meters. 
Orthorectified scenes were mosaiced to produce a single image 
for each one of both provinces. In Tekirdag 44 scenes and 21 
scenes in Agri were mosaiced. The “breakpoint color 
balancing” tool of Erdas Imagine was used to balance the color 
of all images in each province in order to produce a seamless 
color image. 
The control of the visual appearance of the orthoimages was 
done by the optical verification of the images by experienced 
operators. In cases that some areas were identified as 
problematic in terms of visual appearance (as the phenomenon 
of “pulled” / “multiple” idols or the presence of degraded or 
“undulate” forms like “ghosting images”, “wavy features”, 
“smears”), they were further investigated for their possible 
causes and the orthorectifying procedure was executed again. 
The control of the geometry precision of the resulted 
orthoimages was realized with the use of independent control 
points of higher precision than 1/3 of the required final 
precision (2,5m RMS), The coordinates that resulted from the 
measurements of independent control points were compared 
with those that resulted from the measurements of the x,y 
position of the relevant point on the orthoimages with a view to 
calculate the relative statistical figures which determined the 
level of precision of final products. 
3.2. Selection of Reference Parcel: 
Concerning the establishment of LPIS,, different reference 
parcel options have been thoroughly assessed particularly in 
terms of database homogeneity and compulsory use of GIS as 
stipulated by EU legislation. These options are: 
Reference Unit / User 
Based 
Description 
Agricultural Parcel 
One piece of land cultivated by one 
farmer with a single crop or a group 
of crops. 
Farmers Block / JJqJt 
One piece of land cultivated by one 
farmer with one crop or group of 
crops. 
Physical Block 
One piece of land with permanent 
boundaries may contain several crops 
and farmers. 
Reference Unit / Owner 
Based 
Description 
Cadastral Parcel 
Piece of land with specific geometry 
that belongs to physical or legally 
iAentilg 
Figure 1 Definition of reference parcels 
Due to fact that cadastre is concerning property rights, cadastral 
parcels may not correspond directly to the agricultural parcels 
(the location is correct but the boundaries and the area of 
cadastral and corresponding agricultural parcel they do not fit),