Estimating Sky Level

TL;DR

This paper introduces an advanced sky background estimator that uses optimal filtering and statistical modeling to improve accuracy in astronomical images, especially for faint galaxies.

Contribution

It presents a novel sky estimation algorithm that outperforms existing methods in accuracy and robustness, particularly in complex galaxy environments.

Findings

Achieves sky background estimates accurate to 4 ppm.

More robust against galaxy surface brightness profiles.

Better performance compared to traditional sky estimation codes.

Abstract

We develop an improved sky background estimator which employs optimal filters for both spatial and pixel intensity distributions. It incorporates growth of masks around detected objects and a statistical estimate of the flux from undetected faint galaxies in the remaining sky pixels. We test this algorithm for underlying sky estimation and compare its performance with commonly used sky estimation codes on realistic simulations which include detected galaxies, faint undetected galaxies, and sky noise. We then test galaxy surface brightness recovery using GALFIT 3, a galaxy surface brightness profile fitting optimizer, yielding fits to S\'{e}rsic profiles. This enables robust sky background estimates accurate at the 4 parts-per-million level. This background sky estimator is more accurate and is less affected by surface brightness profiles of galaxies and the local image environment compared with other methods.