ft, 
es 
he 
es 
lu- 
ns 
|V- 
ca- 
as, 
o3 
hy- 
99, 
and 
and 
4th 
ngs 
RS, 
ob- 
sed 
hop 
In- 
GEOMETRIC ACCURACY OF HIGH-RESOLUTION DATA 
FOR URBAN PLANNING 
F. YASA', F. Sunar ERBEK?, A. ULUBAY®, C. ÖZKAN* 
‘Istanbul Technical University, Faculty of Civil Engineering, Geomatic Engineering Graduate Program, 34469, Maslak, 
Istanbul, Turkey, yasaf@itu.edu.tr 
“Istanbul Technical University, Faculty of Civil Engineering, Remote Sensing Division, 34469, Maslak, Istanbul, 
Turkey, fsunar@ins.itu.edu.tr 
"General Command of Mapping, 06100, Ankara, Turkey, aulubay@hgk.mil.tr 
^Erciyes University, Geodesy and Photogrammetry Engineering, Remote Sensing Division, 38039, Kayseri, Turkey, 
cozkan(@erciyes.edu.tr 
KEY WORDS: Ikonos, QuickBird, 2D Polynomial Transformations, Geometric Accuracy 
ABSTRACT: 
The very high-resolution satellite data from IKONOS and QuickBird are in a competition to aerial images today. Although the 
IKONOS satellite has provided the world's first source of commercially available high-resolution satellite imagery, QuickBird is 
currently the satellite with the highest resolution for civilian uses. While the IKONOS and QuickBird panchromatic band images 
with 0.82 and 0.61 meter respectively, provide intelligence quality imagery both for military applications, such as monitoring and 
assessment, as well as for civilian applications such as urban planning and mapping. The multi-spectral bands of IKONOS and 
QuickBird multi-spectral images with 3.28 and 2.44 meter respectively provide spectral-radiometric measurements for the many 
applications in land-use mapping, environmental monitoring and resource development. The successful launch of these two high- 
resolution satellites has narrowed the gap between satellite images and aerial photos. In the near future, it could even replace aerial 
photos for some applications depending on the resolution and accuracy requirements. In this study, the geometric accuracy analysis 
has been done with both QuickBird-2003and IKONOS-2002 images for Büyükcekmece region, Istanbul, Turkey. For the accuracy 
assessment, these two high-resolution images were rectified to UTM coordinate system using 1/5000-scaled orthophotos, GPS and, 
hand GPS coordinates. The results obtained were tested at the known points and effectiveness of the use of high-resolution data was 
outlined. 
1. INTRODUCTION 
More than 70 percent of the populations of developed countries 
live in urbanized areas. In developing countries of the third and 
fourth world the migration to urban areas is continuing at an 
increasing rate and tends to rise to a peak of 85 %. Urban spaces 
are the areas developing most dynamically and are characterized 
by an enormous need of information as a basis for planning and 
decision making in order to take measures against the serious 
wide ranging problems arising out of rapid urbanization and 
mostly unplanned development (Krützschmar et al, 2004). 
Traditional land surveying methods, such as field surveys or 
using aerial photogrammetry, are costly and time consuming. In 
order to rapidly derive detail land use information in broad 
areas, it is necessary to use remote sensing techniques. With the 
successful launch of the satellites IKONOS in September of 
1999 and QuickBird in October of 2001 new sources of digital 
imagery are available to decision makers in Provincial, State, 
and Local governments. This improvement in the resolution of 
satellite images has broadened the applications for satellite 
images to areas such as urban planning, data fusion with aerial 
photos and digital terrain models (DTMs), and the integration of 
cartographic features with GIS data. However, these high- 
resolution satellites, such as 1-m resolution IKONOS and 0.6m 
QuickBird, still could not replace the use of aerial photos, which 
have resolution as high as 0.2 to 0.3m. But, these high- 
resolution sensors have narrowed the gap between satellite 
images and aerial photos, permitting the easy updating of urban 
information. Such data the panchromatic resolution from I- 
metre to 70- centimetre and multi-spectral from 4-metre to 2.80- 
metre resolution are likely to stimulate the development of 
urban remote sensing (Fritz, 1999). However, these Very High 
Resolution (VHR) images can not be used directly with map 
base products into a Geographic Information System (GIS) 
because of some distortions. Consequently, multi source data 
integration (vector and raster) for mapping applications requires 
geometric and radiometric correction models. These distortions 
of VHR images are due to radiometric distortions, internal 
sensor geometry, some optical and sensor deficiencies and 
viewing geometry and terrain relief. In addition to these, some 
distortions occur based on the map projection type used. It was 
found that the exact geometric correction of VHR images 
required complex mathematical functions and models, which 
based on principles related to orbitography, photogrammetry, 
geodesy, and cartography (Toutin et al., 2002). For geometric 
correction of VHR images, either 2D/3D non-parametric models 
or rigorous 3D parametric models are being used. Although 2D 
non-parametric models do not take into account sources of 
distortion due to the image formation and terrain relief, they 
have a simple mathematical formation and require fewer 
numbers of GCPs than 3D models. 
In this paper, we investigate a geometric correction procedure 
based on 2D polynomial non-parametric models for Ikonos and 
QuickBird images. We aimed to show the effectiveness of 2D 
polynomial models for VHR images in terms of GCPs obtained 
from different sources. The GCPs were collected by geodesic 
GPS campaing, measuring from photogrammetrically produced 
orthofoto and the hand-GPS. 
1017 
ERA.