Computer vision applications for coronagraphic optical alignment and image processing

TL;DR

This paper explores computer vision techniques like feature extraction, clustering, and search algorithms to improve automated optical alignment and image processing in coronagraphic systems, exemplified by the Gemini Planet Imager.

Contribution

It introduces specific implementations of computer vision methods tailored for coronagraphic optical alignment and calibration tasks, enhancing automation and accuracy.

Findings

Effective feature extraction and clustering for alignment

Successful application of search algorithms for feature detection

Improved automation in optical system calibration

Abstract

Modern coronagraphic systems require very precise alignment between optical components and can benefit greatly from automated image processing. We discuss three techniques commonly employed in the fields of computer vision and image analysis as applied to the Gemini Planet Imager, a new facility instrument for the Gemini South Observatory. We describe how feature extraction and clustering methods can be used to aid in automated system alignment tasks, and also present a search algorithm for finding regular features in science images used for calibration and data processing. Along with discussions of each technique, we present our specific implementation and show results of each one in operation.