Документ взят из кэша поисковой машины. Адрес оригинального документа : http://hea-www.harvard.edu/AstroStat/astro193/ChristopherMerchantz_ay193_fp_slides.pdf Дата изменения: Mon May 4 19:51:06 2015 Дата индексирования: Sun Apr 10 13:44:17 2016 Кодировка: Поисковые слова: п п р р р р р р р р п п п п п п

Testing the Automatic Detection of Directly Imaged Exoplanets Using Wavelet Analysis

Chris Merchantz


Introduction
·

Challenges associated with direct images of planets Distinguish planets in the following way: (1) a minimum flux threshold is from speckles maxima, anddefined for the features that are identified as local (2) the "Exoplanet Recognition Using a Wavelet Analysis Technique"
Fig. 2.--Thick solid line: Example of a normalized spectrum of a planet (yaxis) vs. the number of pixels (x-axis). This spectrum is used as template spectrum to recognize planets in a deep image. Dotted lines: Spectra of a few speckles recognized by the procedure as local maxima and rejected because they did not follow the criteria of selection.

·

·

local maxima are selected as planetar y candidates if their wavelet spectrum coincides with the template spectrum to within a bar of X% extended over the typical FWHM of the PSF. In general, if the error bar is too small, the risk is to not recognize potential planets (especially close to the star). If the error bar is too large, the risk is to recognize as "planets" other sources of noise. In our case X p 10. Our procedure is tested on a deep image of the star V343 Nor (mK p 6.3 mag, d 40 pc, age 12 Myr, and a total exposure time of 22 minutes) obser ved with NACO (Masciadri et al. 2004) in the K s broadband. Along a radial direction (at 20 with respect to the x-axis), a set of six template planets are placed at 0 .2, 0.5, 0 .7, 1 , 1 .5, and 2 from the star with a fixed D M selected by the user (Fig. 1). The image is then filtered from the low spatial frequencies. Two background stars (A and B) are Fig. 3.--Automatic "wavelet analysis" detection of template planets. In each visible at around 2.5 from the central star. In a c(Masciadri & six planets were2004, described in 611, L137) ircular region frame, Raga, simulated as ApJ, Fig. 1 using a planet/star (radius 1 ) around the star, a residual noise is visible. DM calculated following the COND atmospherical models (Baraffe et al. 2003). Top: MP p 3MJ. Bottom: MP p 10MJ. The crosses indicate the positions The automatic recognition of planets in this deep image was of the planets recognized by the automatic procedure. The background stars tested using two sets of template planets having different A and B appear in all of the frames (see text for details). [See the electronic masses, corresponding to different D M with respect to the cenedition of the Journal for a color version of this figure.]


Kalas et al. 2008, Sci, 322, 1345


Kalas et al. 2008, Sci, 322, 1345


Data
· ·

Fomalhaut b (Hubble ACS/HRC, BVI bands) HR 8799bcd (Keck, JHK bands)

Marois et al. 2008, Sci, 322, 1348


Analysis
·

Follow procedure of Masciadri & Raga
· · ·

Convolve image with Mexican hat Locate local maxima Plot spectrum against scale


No. 2, 2004

Results
EXOPLANET RECOGNITION

L139

Normalized Spectrum

HR 8799

Fomalhaut

Fig. 2.--Thick solid line: Example of a normalized spectrum of a planet (yaxis) vs. the number of pixels (x-axis). This spectrum is used as template spectrum to recognize planets in a deep image. Dotted lines: Spectra of a few speckles recognized by the procedure as local maxima and rejected because they did not follow the criteria of selection.

Wavelet Scale

in the following way: (1) a minimum flux threshold is defined


Conclusions
· · · ·

Outlook not optimistic Technique not "automatic" enough Automation not necessary Wavelet analysis for direct imaging reliable only for M > MJup planets at separations ~ 1 as