On-orbit imager installation angles estimation and compensation based on time-series prediction method for GEO optical satellite

The On-orbit geometric calibration of geostationary cameras is literally considered a primary prerequisite for the quantitative applications involving accurate geometric positioning, tracking and identification of space-aeronautics moving targets. Affected by the solar illumination angles and the orbital heat flux of the satellite, the thermal environment around the camera turns out to be of difference in different imaging time and areas, which will cause uncertain spatial thermal deformation among the installation structures, and then change the camera’s geometric positioning model and the final positioning accuracy inevitably. In terms of the problems mentioned above, a novel installation angles estimation method based on Time-Series Prediction Method is proposed to correct the positioning error caused by the spatial thermal deformation of the geostationary satellites. The installation angles estimation algorithm based on Time-Series Prediction Method model is described elaborately. This proposed method overcomes the dependence of conventional methods on the quantity and quality of GCPs, and is able to estimate the angle biases of the installation structure of the camera. Experimental results show that 92.5% and 97.63% of the positioning errors in the testing dataset corrected by the proposed method can be within 2 pixels and 3 pixels, respectively, which is better than that of ± 18 pixels before correction. Generally, this method could be a supplementary to the conventional methods to correct the positioning error caused by spatial thermal deformation of geostationary payloads.