Use of stereo-matching to coregister multiangle data from MISR

The pattern-matching algorithms originally developed for Multi-angle Imaging SpectroRadiometer (MISR) (flying on the Earth Observing System (EOS) Terra platform) cloud retrieval have also proven useful in independently providing quality assurance of the coregistration of multiangle measurements with the nadir view. Two new techniques developed to test the coregistration are described in this paper along with results of the misregistration detection on both historical and current data. No ground-control points are strictly necessary for these calculations—just simultaneous clear-sky land imagery for three cameras and knowledge of the terrain altitude. The difficulty of registration increases with the obliquity of the view angle, so our emphasis is on coregistering to the nadir view. This paper also provides proxy validation of the stereo-matching algorithms for clear-sky land scenes.