Astron. Astrophys. 361, L9-L12 (2000)

3. Reduction procedure

First, standard reduction techniques were applied to the data: bias subtraction, flat-field correction. For each filter, we obtained a set of HD 100546 observations and corresponding PSFs. In spite of the use of a coronograph mask, the disk surface brightness is still dominated by the stellar emission at any distance from the star. In order to retrieve the disk image, one has to remove numerically the starlight wings. The rough substraction of a scaled PSF to HD 100546 images gives poor results because of slight shifts in position on the array between the reference and the object, uncertainties on the fluxes (given by the literature), and a residual background (ADONIS bench emission, different airmasses). We developed a specific method to get an optimum subtraction. For each couple of HD 100546 image () and corresponding , we have to find 3 parameters: a shift (,) between the two images, a scaling factor R, and a residual background Bg. These parameters are estimated by minimizing the following error functional:

where the S function stands for image shift. The sum is performed on a set of pixels (typically 1000) located in a region of the images where no disk emission is expected, and from which non accurate ones are excluded (bad pixels, pixels belonging to area contaminated by diffracted light from the coronograph support). The functional minimum is found using a 0-order minimization algorithm called Powell method (Press 1996). The method was first checked on simulated data whose input parameters (shifts, scaling factor) were recovered with an accuracy better than 5%. In order to evaluate the errors in the subtraction process, test the stability of the PSF, and make sure that the disk obtained is not an artifact created by the process, the same substraction process was applied to the two different PSFs. Fig. 1 shows the resulting images in the J and Ks filters of the disk and the corresponding residuals between the two reference stars. The position of the star under the mask, and therefore the disk center is also carefully determined in this procedure thanks to offset observations (out of the coronograph mask) of the reference stars and the object.

Fig. 1. Upper left: the disk seen in the J band, North is to the top, East to the left. The pixel scale is 0.035 arcsec/pixel. Upper right, the corresponding residuals obtained when applying the reduction method described in the text to two reference stars. Lower frame: same as upper frame but in the Ks filter. The "cross-like" pattern (with fingers orientations at 45,135,225, and 315 degrees) superimposed to the disk emission is produced by astigmatism residuals not corrected by the adaptive optics loop, showing that, contrarily to J band where the correction residuals are still dominated by the seeing, in Ks band under good conditions, one can reach high Strehl ratio values (i.e. system residuals dominating the uncorrected part of the seeing)

© European Southern Observatory (ESO) 2000

Online publication: September 5, 2000
helpdesk.link@springer.de