Navegar por los elementos (1 total)
- Resumen es exacto "In astrophysics and cosmology, as in other areas of physics, the next decade will bring a new generation of extremely powerful instruments, such as Euclid or the Nancy Grace Roman Space Telescope. An accurate model of the point spread function (PSF), i.e., the instrumental response of the optical system, is a fundamental requirement to meet ambitious scientific goals. The WaveDiff model is a semi-parametric wavefront model for data-driven PSF estimation. This method, developed in the CosmoStat lab, builds the PSF model directly on the wavefront space and is based on a differentiable optical model that integrates the physical processes to go from a wavefront error (WFE) to a pixel-level PSF, integrating the wavelength dependency over the instrument bandwidth alongside the spectral energy distribution of observed stars (SED). In this work an error analysis of the model with respect to aberrations on the input SED data is carried out. A method is proposed to alleviate the impact of SED degradation (spectral binning and photometric measurements noise) on the model predictions. The WaveDiff model is optimised and put to test with different dataset sizes to consider the possible benefit of increasing the number of observations used. Finally, a new optimisation paradigm was proposed for the semiparametric model that allows to substantially improve the optimisation of the parametric model decreasing prediction errors both in wavefront and pixel space."
Ordenar por: