26 July 2016 Rotation and translation registration of bandlimited interferometric images using a chirp z-transform
Author Affiliations +
Abstract
Image reconstruction algorithms for wide-field spatio-spectral interferometry require knowledge of registration parameters associated with low-resolution image measurements at various baseline orientations, such that the images can be registered to within the fine resolution of the final desired image. We have developed an image registration procedure that combines a nonlinear optimization algorithm with the sub-pixel precision of chirp z-transform resampling, particularly for rotation and translation, of bandlimited images with non-radially symmetric aberrations. We show the accuracy of this image registration technique on simulated images that have a complexity comparable to scenes observed experimentally with NASA’s wide-field imaging interferometry testbed. Registration to within a tenth of a pixel for translation and within three arcminutes for rotation is demonstrated at the largest simulated noise levels.
© (2016) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Alexander S. Iacchetta, James R. Fienup, David T. Leisawitz, "Rotation and translation registration of bandlimited interferometric images using a chirp z-transform", Proc. SPIE 9907, Optical and Infrared Interferometry and Imaging V, 99073G (26 July 2016); doi: 10.1117/12.2232178; https://doi.org/10.1117/12.2232178
PROCEEDINGS
10 PAGES


SHARE
RELATED CONTENT

Selection of the SPOT5 image compression algorithm
Proceedings of SPIE (October 03 1998)
Imaging algorithm using the bispectrum
Proceedings of SPIE (July 24 1998)
Algorithms for subpixel edge reconstruction
Proceedings of SPIE (June 26 1996)

Back to Top