For researchers focusing on relatively small areas, orthorectification, or image correction that accounts for both imaging geometry and terrain, can be achieved using a relatively straightforward process described in Casana and Cothren (2008), whereby a small image segment is extracted from a larger CORONA scene and then treated as a simple frame camera. Using the focal length and average flying height of the CORONA camera, along with the image size of the segment, the interior orientation of the camera can be estimated as long as the segment represents no more than around 15% of the complete scene. By inputting a series of ground control points (GCPs) derived from other ortho-images or maps, a relatively good orthorectification can be achieved using any number of softcopy photogrammetry software packages, such as ERDAS' Leica Photogrammetry Suite. Our original efforts using this method relied on orthorectified SPOT (Système Pour l'Observation de la Terre) panchromatic imagery for horizontal ground control and SRTM (Shuttle Radar Topography Mission) data for vertical control (Casana and Cothren 2008). In many areas it is rather difficult to find ground control points on SPOT or similar imagery owing to its relatively coarse spatial resolution (10m) and the recent date of the images. Our more recent work has experimented with using Google image data for ground control, with relatively good results. The approach described in Casana and Cothren (2008) is effective over small areas, producing better ortho-images than non-rigorous methods using fewer GCPs. It also preserves the stereo capabilities of images, enabling stereo viewing and extraction of digital topographic data. However, the procedure is time consuming owing to the relatively large number of GCPs that must be collected (15–20 per segment, resulting in no less than 160 for a full CORONA scene). The technique also produces increasing poor results towards the edges of images where distortions are greatest. Given that our goal was to correct hundreds or thousands of images, it was necessary to develop more efficient methods for orthorectification of entire CORONA scenes.
© Internet Archaeology/Author(s)
University of York legal statements | Terms and Conditions
| File last updated: Fri Sept 21 2012