Zustandsschätzung geostationärer Beobachtungssatelliten auf der Basis multispektraler Bildinformationen

Angebote / Angebote:

Geostationary Earth observation weather satellites acquire multispectral images during operational mode using a line scanner. Satellite rotational and translational motion during acquisition phase causes geometrical image distortions. This thesis proposes an alternative approach, which detects those distortions and additionally determines the corresponding geo-position of each pixel. This thesis develops a specific method for geometrical image correction and geo-referencing by evaluating image shift vectors between two images, also known as optical flow. The determined image shift vector field gives a conclusion of the camera motion, which took place during the period of acquisition. Based on the determined optical flow, the satellite's rotational motion is reconstructed. By this way, redundant attitude measurements are obtained from the payload camera as well as the onboard navigation (AOCS). All measurements are fused and filtered by applying a Kalman filter. It estimates the satellite's attitude (primary states) as well as systematic sensor errors (secondary states). A high-precision knowledge of the scanner's line of sight over time is obtained. On this basis the scan imagery is geometrically corrected. At the same time the onboard navigation is supported by the obtained state information.