28 April 2016 Quality metric in matched Laplacian of Gaussian response domain for blind adaptive optics image deconvolution
Author Affiliations +
Abstract
Adaptive optics (AO) in conjunction with subsequent postprocessing techniques have obviously improved the resolution of turbulence-degraded images in ground-based astronomical observations or artificial space objects detection and identification. However, important tasks involved in AO image postprocessing, such as frame selection, stopping iterative deconvolution, and algorithm comparison, commonly need manual intervention and cannot be performed automatically due to a lack of widely agreed on image quality metrics. In this work, based on the Laplacian of Gaussian (LoG) local contrast feature detection operator, we propose a LoG domain matching operation to perceive effective and universal image quality statistics. Further, we extract two no-reference quality assessment indices in the matched LoG domain that can be used for a variety of postprocessing tasks. Three typical space object images with distinct structural features are tested to verify the consistency of the proposed metric with perceptual image quality through subjective evaluation.
© 2016 Society of Photo-Optical Instrumentation Engineers (SPIE)
Shiping Guo, Rongzhi Zhang, Yikang Yang, Rong Xu, Changhai Liu, Jisheng Li, "Quality metric in matched Laplacian of Gaussian response domain for blind adaptive optics image deconvolution," Optical Engineering 55(4), 043108 (28 April 2016). https://doi.org/10.1117/1.OE.55.4.043108 . Submission:
JOURNAL ARTICLE
8 PAGES


SHARE
Back to Top