GEOMETRIC CORRECTION AND OBJECT EXTRACTION OF IKONOS IMAGES IN THE APPLICATIO OF MAP REVISION
| Category | GIS & Remote sensing |
|---|---|
| Group | GSI.IR |
| Location | Geomaics conference 86 |
| Holding Date | 28 January 2008 |
GEOMETRIC CORRECTION AND OBJECT EXTRACTION OF IKONOS IMAGES IN THE APPLICATIO OF MAP REVISION
[ Mohammad Sohrabinia ] - Shahid Beheshti Uiversity, Faculty of Earth Scrinces, Department of Remote Sensing and GIS, Tehran, Iran
[ Saeid Sadeghian ] - Geomatics College of National Cartographic Center (NCC), Tehran , Iran
[ Ali Akbar Matkan ] - Shahid Beheshti Uiversity, Faculty of Earth Scrinces, Department of Remote Sensing and GIS, Tehran, Iran
[ Dadfar Manavi ] - Nationl Cartographic Center (NCC), Tehran,
abstract
In this study , we have used different geometric correction methods to achieve a sub-pixel accuracy for the IKONOS image. After achieving this goal, we have produced an ortho – image of the panchromatic band and the multispectral image. Using these corrected image, we have produced a pan – sharpened image. Afterwards, we discuss information content and object extraction from IKONOS satellite image in order to update large scale urban maps.
At first discuss the best methods of geometric correction and pan-sharpening of IKOKOS images. While the panchromatic band has one – meter pixel size, achieving and accuracy of less than one meter will be essential. Geometric corrections of IKONOS Geo image with digital maps and some Ground Control Points (GCPs) had been implemented. We obtained 0.4 meter accuracy for Independet Check Points (ICPs) with polynomial and rational models, which is satisfactory subpixel accuracy. After achieving the ultimate sub-pixel precision, extracting information from the image and revising digital maps has been analyzed. Visual interpretation , automatic classification and fuzzy object extraction methods were analyzed to produce new vector maps. Together with the image itself, these new maps were used for change detection through the overlay and comparison of the old maps. At last, we analyzed different feature classes of 1:10000, 1:5000 and 1:2000 scale maps and compared them to the information content on an IKONOS pan – sharpened image.
keywords
Spectral resolution , spatial resolution , mathematical modeking , geometric correction , image fusion , object extraction , information content , revision