The Point Spread Function Reconstruction by Using Moffatlets - I

TL;DR

This paper compares three PSF reconstruction methods—Gaussianlets, Moffatlets, and EMPCA—using simulated LSST star images, finding Moffatlets outperform Gaussianlets and EMPCA offers flexibility but with noise sensitivity.

Contribution

The study introduces and evaluates three PSF reconstruction methods, highlighting the superior performance of Moffatlets in weak lensing surveys.

Findings

Moffatlets outperform Gaussianlets in PSF reconstruction.

EMPCA is more flexible but sensitive to noise.

Moffatlets maintain PSF size and ellipticity well.

Abstract

The shear measurement is a crucial task in the current and the future weak lensing survey projects. And the reconstruction of the point spread function(PSF) is one of the essential steps. In this work, we present three different methods, including Gaussianlets, Moffatlets and EMPCA to quantify their efficiency on PSF reconstruction using four sets of simulated LSST star images. Gaussianlets and Moffatlets are two different sets of basis functions whose profiles are based on Gaussian and Moffat functions respectively. Expectation Maximization(EM) PCA is a statistical method performing iterative procedure to find principal components of an ensemble of star images. Our tests show that: 1) Moffatlets always perform better than Gaussianlets. 2) EMPCA is more compact and flexible, but the noise existing in the Principal Components (PCs) will contaminate the size and ellipticity of PSF while Moffatlets keeps them very well.