Model driven-based extended target piston sensing for optical synthesis aperture imaging system

The next generation of Earth observation on-orbit optical telescopes will provide high-resolution imaging of complex scene by using the aperture synthesis technique. Fast correction of pistons among each subaperture is the critical factor to preserve the synthetic image quality. Numerous point-spread function (PSF)-based methods have been proposed for piston detection, requiring a single star as a calibration source. When the exploration mission is similar as geological hazard monitoring and terrain landform analysis, the interference fringes that can characterize the piston will be submerged in the feature information of the complex scene, making it difficult to directly extract the pistons from the image. In this paper, by setting the non-redundant exit-pupil distribution, we innovatively construct the mathematical model of the eigenvector for extend target piston sensing, termed the frequency secondary-peak piston extraction (FSPE). Based on the polynomial fitting of piston response eigenvector, the high-precision real-time piston correction is achieved through the inverse solution of several quintic functions. The proposed method can directly use images captured by CCD detector for piston sensing without a calibration source and any other additional optics. Considering the simple operation and real-time performance of this method, we believe that it can further promote the application of synthetic aperture telescope for high-resolution Earth observation.