Authors

Dmitry Savransky, Tyler D. Groff, N. Jeremy Kasdin

Affiliations

Department of Mechanical and Aerospace Engineering, Princeton University

Abstract

We evaluate the feasibility of applying Bayesian detection techniques to discovering exoplanets using high contrast laboratory data with simulated planetary signals.  Background images are generated at the Princeton High Contrast Imaging Lab (HCIL), with a coronagraphic system utilizing a shaped pupil and two deformable mirrors (DMs) in series. Estimates of the electric field at the science camera are used to correct for quasi-static speckle and produce symmetric high contrast dark regions in the image plane.  Planetary signals are added in software, or via a physical star-planet simulator which adds a second off-axis point source before the coronagraph with a beam recombiner, calibrated to a fixed contrast level relative to the source.  We produce a variety of images, with varying integration times and simulated planetary brightness. We then apply automated detection algorithms such as matched filtering to attempt to extract the planetary signals. This allows us to evaluate the efficiency of these techniques in detecting planets in a high noise regime and eliminating false positives, as well as to test existing algorithms for calculating the required integration times for these techniques to be applicable.