12 September 2012 Toward 5D image reconstruction for optical interferometry
Author Affiliations +
Abstract
We report on our progress toward a flexible image reconstruction software for optical interferometry capable of "5D imaging" of stellar surfaces. 5D imaging is here defined as the capability to image directly one or several stars in three dimensions, with both the time and wavelength dependencies taken into account during the reconstruction process. Our algorithm makes use of the Healpix (Gorski et al., 2005) sphere partition scheme to tesselate the stellar surface, 3D Open Graphics Language (OpenGL) to model the spheroid geometry, and the Open Compute Language (OpenCL) framework for all other computations. We use the Monte Carlo Markov Chain software SQUEEZE to solve the image reconstruction problem on the surfaces of these stars. Finally, the Compressed Sensing and Bayesian Evidence paradigms are employed to determine the best regularization for spotted stars. Our algorithm makes use of the Healpix (reference needed) sphere partition scheme to tesselate the stellar surface, 3D Open Graphics Language (OpenGL) to model the spheroid, and the Open Compute Language (OpenCL) framework to model the Roche gravitational potential equation.
© (2012) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Fabien Baron, Fabien Baron, Brian Kloppenborg, Brian Kloppenborg, John Monnier, John Monnier, } "Toward 5D image reconstruction for optical interferometry", Proc. SPIE 8445, Optical and Infrared Interferometry III, 84451D (12 September 2012); doi: 10.1117/12.926274; https://doi.org/10.1117/12.926274
PROCEEDINGS
9 PAGES


SHARE
Back to Top