Exoplanet Recognition Using a Wavelet Analysis Technique

The recognition of a planet orbiting around a star (and in general of a faint companion in a binary system) in direct imaging observations is limited at distances smaller than around 1'' by the speckle and photon noise. There are also other sources of noises as artifacts produced by the adaptive optics (AO) systems that make the identification of a planet even more difficult. Techniques to recognize a planet in an automatic (and objective) way are therefore quite useful. In this Letter, we test the ability of a "wavelet analysis" technique for recognizing exoplanets in deep images obtained with ground-based telescopes and AO facilities. We present the automatic procedure developed for the detection of exoplanets, and we validate it on a deep image of a young, nearby star, which has been one of the targets of a planet search survey done with NACO/VLT. Preliminary tests show that with the proposed algorithm, it is indeed possible to recognize, in an automatic way, a planet of 12MJ at a distance of 05, a planet of 5MJ at a distance of 07, and a planet of 3MJ at a distance of 1'' from the star. We also show that a star/planet systems with a maximum ΔM = 4.7 mag contrast in the K band at 02 can be recognized automatically. This is a typical contrast for a 5MJ-10MJ planet and a late-type star (late M spectral type).