Research On Object Positioning With SPOT 5 Satellite Image Under Lack Of Ground Control Points

The image positioning plays an important role in the process of photogrammetry. But the precision of image positioning has been greatly relying on the precision of Ground Control Points(GCPs) for a long period. The number and distribution of the GCPs directly influences the precision of the object positioning on the ground. However, it is very difficult to obtain the precise position coordinates of the GCPs in the areas such as deserts, oceans or foreign lands. In this paper, we have a research on the model of direct positioning in SPOT 5 images without supports of any GCPs and positioning based on Rational Function Model (RFM) as well as a positioning technology with auto matched linking points to extrapolating in SPOT 5 images using a few GCPs. This paper shows the following works we have done:Firstly, the changing rules of satellite ephemeris data and its pose data are analyzed. On the basis of this former work, the values of satellite ephemeris and its pose at the time of any scanned line in the image are worked out by using a proper model of inserting.Secondly, the technology of SPOT 5 HRG single image positioning and HRS stereo image positioning under no GCPs are studied and practiced. After the experiment, the results of both ways are analyzed, and the rules and their influenced factors are found out. According to the results, we presented a new method to improving the image positioning precision of HRG under no GCPs by sub-area bilinear inserting method to evaluate h. By the new way, it improved the elevation precision greatly.Thirdly, the model of SPOT 5 HRS image positioning based on RFM under no GCPs is researched and practiced. We got the values of exterior orientation elements in any scanned line in the image from satellite ephemeris data and its pose data. Following the work done above, a virtual controlled grid with strict geometry model is created which instead GCPs to accomplish calculation of RFM. At last, positioning in stereo images of SPOT 5 HRS under no GCPs is achieved.Finally, several ways of some familiar orbital extrapolating method are analyzed. On the basis of researching and practicing auto matching between images of HRS, we proposed a new plan of extrapolating position using a few GCPs and public tie points gained by the auto matching between some pairs of stereo images of SPOT 5 HRS. According to the experiments, the plan was validated in a certain extent.