[en] The purpose of the presented research project is to investigate how present-day EO-technology can support local and regional decision-making, particularly in Belgium, and to develop prototype versions of value-added products that fulfill some of the actual information needs, as expressed by Belgian authorities at the local and at the regional level. One of the technical research modules is focussing on the geometrical aspects of VHR data processing.
The geometric precision of VHR data (QuickBird, IKONOS) is systematically evaluated, more specific for an urban and suburban environment using the GCP information obtained by GPS measurements. The 2D-georeferenced images is compared with ortho-rectified images in terms of their geometric accuracy. This will provide useful information about the limitations of 2Dcorrected data, and the necessity of acquiring ortho-rectified data, for different kinds of applications in each of the studied areas. The oblique viewing angle of images produced by VHR sensors like IKONOS and QuickBird has a strong influence on the geometrical displacement of objects, and on the occurrence of hidden areas in the image. The degree of displacement and the percentage of hidden areas are determined by the interaction between viewing angle, position of the sun, object height, and terrain characteristics (slope, aspect). In other words, the range of viewing angles that can be considered acceptable for purchasing an image of a certain area will depend on the morphological characteristics of the area. To study the interaction between all parameters involved, a theoretical model is implemented. First, the model is used to systematically study the impact of viewing angle on object displacement, shadow length and size of hidden areas in an image, by simultaneously varying the different morphological parameters. Next, the impact of viewing angle on the geometry of the image is simulated using the theoretical model and a DSM from aerial photography.
The geometrical accuracy of a DSM from VHR data is compared with the accuracy of the DSM, derived from aerial photography at scales of 1:12000 and 1:4000). The satellite-derived DSM is used for ortho-rectification of VHR imagery. The accuracy of the ortho-rectified images is compared with ortho-rectifications of the same images, based on the aerial photo DSM.
