Satellite attitude instability effects on stereo images

This paper describes an original method to identify microvibrations of satellite platforms using only a pair of stereo images. Image pairs contain a twofold information: the landscape details and the platforrn microvibration corruptions (roll, pitch and yaw). We propose identification methods in the case of a push-broom acquisition mode with a 6000 pixel CCD linear array (in the case of the French satellite SPOT). This method relies on the image acquisition principle, on the nature of the observed landscape and finally on some concepts of surface geometry. The algorithms providing such identification have been implemented and applied to simulated data, based on a true digital model terrain of Marseille (France). The results of our vibration identification method are presented here. Comparisons with traditional methods are also presented in this paper.