PRECISE GEOREFERENCING IN THE ABSENCE OF GROUND CONTROL: A STRIP ADJUSTMENT APPROACH

摘要

The options available for orientation of satellite imagery in cases requiring optimal metric precision are generally rational functions, where the RPCs are provided by the image supplier, or a rigorous physical model. In either case, it is ground control that facilitates pixel-level georeferencing via a compensation for systematic errors in the sensor metadata. In many remote parts of the world, where satellites are the most viable source of imagery for mapping, there is a lack of ground control, which thus precludes bias-free georeferencing. This paper reports on a practical means of overcoming this problem, namely the orientation of long strips of imagery through a bundle adjustment process that requires ground control at the endpoints only. The adjustment utilises a rigorous sensor orientation model. RPCs are then generated for each scene within the strip of images from the adjusted sensor orientation data. These facilitate bias-free georeferencing without reference to ground control. The approach discussed has previously proven successful for automated orthoimage generation from ALOS imagery, and in this paper its application to a 7-scene strip of QuickBird imagery is reported.