Georeferencing from Geoeye-1 Imagery: Early Indications of Metric Performance

GeoEye-1 was launched in September 2008, and after some five months of commissioning commenced commercial operations in February, 2009. With its 0.41m panchromatic and 1.65m multispectral resolution, GeoEye-1 represents a further step along the way to higher resolution capabilities for remote sensing satellites. Given experiences with precise georeferencing from its sister satellite, IKONOS, initial accuracy expectations for 3D georeferencing with ground control entered the 20-40cm range, and geolocation utilising metadata (orbit and attitude recordings) alone is specified at 2-3m. This paper describes an early experimental assessment of the accuracy of georeferencing from GeoEye-1 imagery. A stereo panchromatic image pair covering the Hobart HRSI test field in Australia was utilised in the testing. This test field, originally established to support metric testing of IKONOS imagery, comprises more than 100 precisely measured GCPs, of which 55 were deemed suitable for the GeoEye-1 tests. Three aspects were investigated with the resampled 50 cm imagery: the first was the geolocation accuracy attainable from utilising vendor supplied RPCs, ie those generated utilising metadata alone; the second was the accuracy attainable via bias-corrected RPCs; and the third involved application of a linear empirical model, not so much to offer an alternative geopositioning capability, but more to gain an insight into the degree of linearity of GeoEye-1’s east-to-west image scanning. The paper will highlight the fact that with bias-corrected RPCs and a single GCP, the RMS georeferencing accuracy reaches the unprecedented level of 0.10m (0.2 pixel) in planimetry and 0.25m (0.5 pixel) in height.