Georeferencing Accuracy of Geoeye-1 Stereo Imagery: Experiences in a Japanese Test Field

High-resolution satellite imagery (HRSI) is being increasingly employed for large-scale topographic mapping, and especially for geodatabase updating. As the spatial resolution of HRSI sensors increases, so the potential georeferencing accuracy also improves. However, accuracy is not a function of spatial resolution alone, as it is also dependent upon radiometric image quality, the dynamics of the image scanning, and the fidelity of the sensor orientation model, both directly from orbit and attitude observations and indirectly from ground control points (GCPs). Users might anticipate accuracies of, say, 1 pixel in planimetry and 1-3 pixels in height when using GCPs. However, there are practical and in some cases administrative/legal imperatives for the georeferencing accuracy of HRSI systems to be quantified through well controlled tests. This paper discusses an investigation into the georeferencing accuracy attainable from the GeoEye-1 satellite, and specifically the 3D accuracy achievable from stereo imagery. Both direct georeferencing via supplied RPCs and indirect georeferencing via ground control and bias-corrected RPCs were examined for a stereo pair of pansharpened GeoEye-1 Basic images covering the Tsukuba Test Field in Japan, which contains more than 100 precisely surveyed and image identifiable GCPs. Salient aspects of the investigation are discussed, including aspects of sensor orientation via bias-corrected RPCs, whether GCPs are required, and the relationship between geolocation accuracy and the number and quality of GCPs. The paper discusses the results obtained, which indicated that the direct georeferencing accuracy obtained was within that specified for GeoEye-1, namely a 2m Circular Error 90% (CE90) in planimetry and a 3m Linear Error 90% (LE90) in height. The use of a few GCPs improved geopositioning accuracy to around 0.35m (0.7 pixel) in planimetry and 0.7m (1.4 pixel) in height. * Corresponding author.