A problem with coded-mask telescopes is the achievable angular resolution. For example, with the standard cross-correlation (CC) analysis, the INTEGRAL IBIS/ISGRI angular resolution is about 13’. We are currently investigating an iterative Lucy-Richardson (LR) algorithm. The LR algorithm can be used effectively when the PSF is known, but little or no information is available for the noise. This algorithm maximizes the probability of the restored image, under the assumption that the noise is Poisson distributed, which is appropriate for photon noise in the data, and converges to the maximum likelihood solution. We have modified the classical LR algorithm, adding non-negative constraints. It doesn’t take into account of the features leading to a difference in PSF depending on position in the field of view (dead pixels, gaps between modules etc), which are easily corrected for in the classical CC analysis, so we must correct for these either after the restoration of the image or by modifing the data before the sky reconstruction. We present some results using real IBIS data indicating the power of the proposed reconstruction algorithm.
Improving the angular resolution of coded aperture instruments using a modified Lucy-Richardson algorithm for deconvolution
SAMBO, Lara;BONETTINI, Silvia;ZANGHIRATI, Gaetano;FRONTERA, Filippo
2009
Abstract
A problem with coded-mask telescopes is the achievable angular resolution. For example, with the standard cross-correlation (CC) analysis, the INTEGRAL IBIS/ISGRI angular resolution is about 13’. We are currently investigating an iterative Lucy-Richardson (LR) algorithm. The LR algorithm can be used effectively when the PSF is known, but little or no information is available for the noise. This algorithm maximizes the probability of the restored image, under the assumption that the noise is Poisson distributed, which is appropriate for photon noise in the data, and converges to the maximum likelihood solution. We have modified the classical LR algorithm, adding non-negative constraints. It doesn’t take into account of the features leading to a difference in PSF depending on position in the field of view (dead pixels, gaps between modules etc), which are easily corrected for in the classical CC analysis, so we must correct for these either after the restoration of the image or by modifing the data before the sky reconstruction. We present some results using real IBIS data indicating the power of the proposed reconstruction algorithm.I documenti in SFERA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
