Users will receive pre-processed data from CFHT. Terapix can provide fully astrometrically and photometrically calibrated stacked images derived from these pre-processed images.
Building on experience acquired with MegaCam processing, two points need particular attention. Although the WIRCam field of view is smaller than MegaCam (meaning that higher-order terms are less important in the astrometric solution), careful attention must be paid to the quality of the input catalogues, as the cosmetic quality of the WIRCam detectors is much lower than MegaCam. In addition, infrared observations necessitate a second order analysis for background subtraction, because of stronger emission of faint objects in these wavelengths.
Tools developed at Terapix are used for WIRCam data reduction, which comprises the following four steps:
Quality assessment and weight maps production
During this step general properties of images are computed: FWHM, background level, PSF quality, galaxy and stellar counts. Catalogues for astrometric and photometric calibrations are extracted and weight maps generated for every image. Particular attention is given to the separation between spurious and real detections, because of the non-optimal cosmetic properties of the WIRCam detectors. Production of "cleaned" catalogues is critical for deriving a reliable astrometric solutions. In some cases CFHT provides cubes of images. Infrared observations are often performed using dithering or micro-dithering strategy. This technique produce a lot of images with small offsets around a central position. Archiving reasons make cubes of i mages preferable because for more images only one header is used. After some tests, in particular about the computation of the astrometric solution and the stacks production, it seems that to keep the difference between single images or cubes during the process was unnecessary. Every slice is actually an image with changing seeing, sky features, photometric quality and zero-points. Hence, all data process is then performed on images having NAXIS = 2.
Precise astrometric and photometric calibrations
A precise astrometric calibration is required to resample the input images. In particular WIRCam observations can also be carried out with an optional micro-dithering observation strategy. This technique allows, in principle, to construct composite images with smaller pixel scale, providing the pixel offsets are precisely known. To perform a correct and general resampling of images and avoid the problem of non-accurate offsets, we don't distinguish between processing of micro-dithered or non-micro-dithered images. For this reason producing a clean input catalogue for the astrometric solution is particularly important. The location of bad pixels and detector defects is fixed and they are easily mismatched as objects if observations have small offsets.
The astrometric solution is calculated using Scamp which also provides relative photometric flux scaling between images and detectors.
Images are resampled and combined using Swarp. An automatic procedure providing optimal "two-pass" background subtraction for infrared images has been developed. In infrared observations the background is affected by faint objects in general not detected in single images but present if stacks are deep enough. When the background is computed in the single image before combination, it is in general overestimated, affecting final magnitudes that are consequently underestimated. Therefore, a preliminary stack is produced, objects are identified on it and background estimation on single images is remade masking the objects themselves. Then a second and final stack is produced.
Catalogues and final quality assessment delivery.
Quality assessment on the final stack is performed producing seeing and background estimations and a preliminary star galaxy separation. All these results are presented as a HTML page. A catalogue is also generated, containing basic object parameters which is subsequently used for the quality assessment procedures. If observations in several filters have been taken, a chi2-image can also be provided with catalogs containing magnitudes in each filter for all objects.