TWO-DIMENSIONAL RECTIFICATION TECHNIQUES FOR QUICKBIRD SATELLITE IMAGES

Abstract

1.1 Statement of the problem

Every satellite image has some distortions that affect the geometric accuracy of these images; these distortions are classed in two groups, random and systematic distortions. The systematic distortions are corrected by applying formulas derived by modeling the source of the distortions mathematically, as an example of systematic distortions is involved in multispectral scanning from satellite altitudes is the eastward rotation of the earth beneath the satellite during imaging. Random distortions, and residual unknown systematic distortions are corrected by analyzing well-distributed ground control points (GCPs) occurring in an image. The satellite images are delivered after applying some correction processes like radiometric distortions, but some of the images must be rectified to remove other types of distortions, so rectification and orthorectification processes are applied. If the area that is covered with the image is flat, it is enough to apply the rectification process (which will be explained in section 3.3), but if this area contains relatively high altitude variations, orthorectification process (which also will be explained in section 3.4) should be applied to remove the relief displacement distortions. To apply the rectification and orthorectification processes, it is necessary to acquire the fundamental model of the satellite, this model relates between the point on the image and its conjugate on the earth. The fundamental model gives the best accuracy, but some types of images are not provided with the fundamental model, so it is necessary to get alternative models for the performance of the rectification and orthorectification processes. Some types of the alternative models are, polynomial models, projective model, affine model, Helmer! model, rational polynomial models, and other models, ( the fundamental and alternative models will be explained in section 2.3). For this research, two QuickBird standard images for Fayoum city in Egypt, were used, the P2 and P3 images cover an area in the middle strip of collected The second operation that was applied on the Fayoum QuickBird standard images was the orthorectification process, this process is necessary because the Fayoum city contains relatively high altitude variations, and the orthorectification process improves spatial accuracy. For the orthorectification process, the Intergraph Image Analyst was used. Since the second order polynomial gave the best results in the rectification process, it was also used in the orthorectification process. The orthorectification process was applied on the previous P2 and P3 images, and also by using the observations (on the images P2 and P3) of different two operators. It was found that the spatial accuracy that resulted from the orthorectification process was from 0.80 to 0.90 meters. So when the AOl contains relatively high altitude variations, the orthorectification operation is recommended, because this operation removes the distortions that are resulted from sensor tilt and relief displacements, this in tum improves the spatial accuracy of the satellite image. Also, the effect of GCPs number on resulted accuracy was studied. The orthorectification operation was repeated by reducing the number of GCPs gradually in the two images. It was found that the accuracy of orthorectification can be in the previous rage by using 20 control points. By comparing the results of the rectification and orthorectification processes for the images P2 and P3 is was found that the spatial accuracy of the rectification process for the P2 image was not improved remarkably, but in the P3 image their was a remarkable improvement on the accuracy of rectification process after the performance of the orthorectification process. These results returned to that the area covered by P2 is flat in its overall parts, but the area covered by P3 image contains variations in altitudes.