Restoration of observed image with an unknown space-variant blur from wide-field telescope

The field-of-view (FOV) of wide-field telescope is mostly beyond the scope of isoplanatic angle. Though the aberrated wave-front within a limited range close to guide star could be corrected accurately by adaptive-optics (AO) system, the image quality of object region deviation from guide star is dropped severely. The post-processing technique to restore degraded image observed from wide-field telescope is a good compensation for the limit of AO system. Restoration method of space-variant point spread function (PSF) degraded image which are used to overcome the turbulence-induced anisoplanatic effect, can improve image quality in the whole field of view of wide-field telescope, and restore image resolution to optical diffraction-limited level in free space. A multi-frame blind deconvolution restoration algorithm based on image sectioning method is presented here. The image is divided into several subimages, where the PSF of each subimage is assumed to be space-invariant. The conjugate gradient optimization algorithm based on maximum likelihood estimation (MLE) in space domain is adopted to estimate the space-invariant PSF and object information in each block. In order to reduce blocking artifacts at the subregion boundaries, larger, overlapping subregions as well as the Mumford- Shah regularization are used, and then the restored sections are extracted from their center. The resolution of the restored image using 5 satellite model blurred image frames through image sectioning method is enhanced significantly, compared to the space-invariant PSF restoration approach. There are more details about the original object, and it shows the proposed algorithm is valid.