In the Spirit of Lyot
Exoplanets detection by direct imaging remains one of the most challenging field of astronomy. The very high contrast between the host star and its orbiting companions can prevent the exoplanets detection in a single dataset. However, combining the information of several observations of the same targeted star taken at different epochs can boost the detection limits to unprecedented levels. We describe a new algorithm named PACOME that optimally combines multi-epoch datasets to improve the exoplanets detection limits while taking into account their orbital motion. Our method is based on the PACO algorithm that learns the stellar contamination from the data spatial correlations to estimate unbiased signal-to-noise ratios (SNR) of the brightness of potential orbiting companions. From these unique PACO outputs, PACOME computes a large number of Keplerian orbits and evaluates their likelihood by computing the associated combined SNR. The presence of a potential exoplanet can be assessed by the orbital parameters maximising the combined SNR. PACOME benefits from PACO’s high sensitivity and is faster than standard multi-epoch algorithms. We demonstrate its efficiency using SPHERE-IRDIS datasets from two exoplanetary systems (HR 8799 and HD 95086). For these systems we demonstrate unprecedented boost in the combined SNR of the known exoplanets and an accurate estimation of their orbital parameters compatible with the literature. New companions to HR 8799 and HD 95086 were not detected in these datasets. In the future, we will adapt PACOME to process ASDI data using SPHERE- IFS in order to increase the detection limits further and possibly detect new fainter exoplanets.