LSST and the Dark Sector: Image Processing Challenges

TL;DR

This paper introduces a 'Multi-Fit' algorithm for analyzing dithered galaxy exposures, improving the accuracy of weak lensing measurements by fitting individual images to reduce systematic errors and increase usable galaxy counts.

Contribution

The paper presents a novel 'Multi-Fit' method that directly fits individual exposures, enhancing weak lensing analysis beyond traditional stacking techniques.

Findings

Increases the number of usable small, high-redshift galaxies.

Reduces systematic shear errors in weak lensing measurements.

Improves the accuracy of dark matter and dark energy studies.

Abstract

Next generation probes of dark matter and dark energy require high precision reconstruction of faint galaxy shapes from hundreds of dithered exposures. Current practice is to stack the images. While valuable for many applications, this stack is a highly compressed version of the data. Future weak lensing studies will require analysis of the full dataset using the stack and its associated catalog only as a starting point. We describe a "Multi-Fit" algorithm which simultaneously fits individual galaxy exposures to a common profile model convolved with each exposure's point spread function at that position in the image. This technique leads to an enhancement of the number of usable small galaxies at high redshift and, more significantly, a decrease in systematic shear error.