Abstract
In order to achieve higher resolutions, current earth-observation satellites use larger lightweight main mirrors which are usually deformed over time, impacting on image quality. In the context of active optics, we studied the problem of correcting this main mirror by performing wavefront estimation in a closed loop environment. To this end, a Shack-Hartman wavefront sensor (SHWFS) used on extended scenes could measure the incoming wavefront. The performance of the SHWFS on extended scenes depends entirely on the accuracy of the shift estimation algorithm employed, which should be fast enough to be executed on-board. In this paper we specifically deal with the problem of fast accurate shift estimation in this context. We propose a new algorithm, based on the global optical flow method, that estimates the shifts in linear time. In our experiments, our method proved to be more accurate and stable, as well as less sensitive to noise than all current state-of-the-art methods.
Highlights
Adaptive optics was originally developed for the field of astronomy to remove image aberrations induced by wavefronts propagating through Earth’s atmosphere
In astronomical imaging a wavefront sensing device is frequently used in conjunction with a deformable mirror in order to correct the undesired effects of atmospheric turbulence, improving the quality of sensed images
It uses an array of lenslets to measure the deformation of the incoming wavefront
Summary
Adaptive optics was originally developed for the field of astronomy to remove image aberrations induced by wavefronts propagating through Earth’s atmosphere. Michau et al [6] was among the first to propose an experimental implementation for using a Shack-Hartmann sensor on extended sources They compute the discrete cross-correlation of the two images, and estimate the subpixel shift by taking the centroid of the shifts with higher correlation value. Sidick et al [9] proposed ACC (Adaptive Cross-Correlation), an iterative method that estimates the shift in the Fourier domain by phase correlation [12]. The second image is resampled (in the frequency domain) by the current shift estimate This method attains very low errors of the order of 0.05 pixels by using fewer than 6 iterations.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
