Beam Profiling with Noise Reduction From Computer Vision and Principal Component Analysis for the MAGIS-100 Experiment

TL;DR

This paper introduces a novel, cost-effective beam profiling method using computer vision and principal component analysis to accurately characterize laser beams in the MAGIS-100 atom interferometer, reducing systematic errors.

Contribution

It presents a new beam profiling technique combining CMOS imaging, computer vision, and PCA to improve laser beam characterization in quantum sensors.

Findings

Effective noise reduction in beam profiles

Accurate characterization of laser aberrations

Cost-efficient and adaptable method

Abstract

MAGIS-100 is a long-baseline atom interferometer that operates as a quantum sensor. It will search for dark matter, probe fundamental quantum science, and serve as a prototype gravitational wave detector in the 0.3 to 3~Hz frequency range. The experiment uses light-pulse atom interferometry where pulses of light create the atom optics equivalents of beamsplitters and mirrors. Laser beam aberrations are a key source of systematic error for MAGIS-100, and accurately characterizing the laser beam spatial profile is therefore essential. In this paper, we describe a new and efficient beam profiling technique. We use a low-cost CMOS camera affixed to a translating and rotating optomechanical mount to image the beam, then employ computer vision and principal component analysis to minimize background noise and produce accurate beam profiles for a laser incident on a variety of aberration-inducing optical elements.