Mapping of urban features (e.g. roads and buildings) from satellite images has gained enormous research interest. Geoeye-1 allows for accurate and reliable extraction and characterization of even more details of the earth surface. Large scale maps could be achieved from both Panchromatic (PAN) Geoeye-1 (0.5 m resolution) and MultiSpectral (MS) Geoeye-1 (2 m resolution) images. In order to exploit the maximum capabilities of the PAN and MS images, PAN-sharpened product has been generated. This research deals with exploring the large scale mapping capability of PAN, MS and fused Geoeye -1. Rectification of PAN image (flat terrain) using three rectification methods has been performed, results indicate that the second order polynomial is the best followed by the third order polynomial then first order polynomial model. In general the RMS decrease with increasing of number of CPs. Results also indicate that for the selected study area there is no need for higher order polynomial, the first order polynomial is enough because it gives satisfied results. In this research three different image fusion methods modified Intensity-Hue-Saturation( modified IHS) , Principal Component Analysis( PCA), and Gram-Shmidt (GS) have been used to fuse two co-registered optical images of the same sensor, PAN and MS Geoeye-1, to improve the interpretation. The fused images obtained for Geoeye-1 (PAN& MS) were compared and analyzed visually and statistically for urban mapping. Overall the method of Gram-Shmidt was found to be the most accurate and objective when the results were compared with the PCA and modified IHS methods. The usefulness of the fused images in urban feature extraction have been analyzed and compared to PAN and MS images. Results proved that the planimetric accuracy of PAN Geoeye-1 rectified image met the 1:2500 Map Accuracy Standards. Also from the assessment of the information content of the PAN Geoeye-1 image, it was found that it is consistent with the map scale of 1:2500. So it is possible to produce the same scale with better interpretation using the fused image. The planimetric accuracy of multispectral Geoeye-1 rectified image met the 1:5000 National Map Accuracy Standards. Also from the assessment of the information content of the multispectral Geoeye-1 image, it was found that it is consistent with the map scale of 1: 5000. |
