Image detrending

The KiDS-ESO-DR2 pipeline is based on the Astro-WISE optical pipeline described in McFarland et al. (2013, ExA 35, 79) (MF13). On this and the following pages we summarize the processing steps and list KiDS-specific information (i.e., KiDS process configuration and departures from the Astro-WISE optical pipeline).

Detrending of the raw data consists of the following steps.

  • Cross-talk correction. Electronic crosstalk occurs between CCDs #93, #94, #95 and #96, resulting in faint imprints of bright sources on neighbouring CCDs. A correction was made for crosstalk between CCDs #95 and #96, where it is strongest (up to 0.7%).
  • De-biasing and overscan correction. First, for each science and calibration exposure the overscan is subtracted per row (using method 6, see MF13). Second, a daily overscan-subtracted bias is subtracted. This is a daily average of 10 biases with 3σ-rejection.
  • Flat-fielding. A single masterflat (per CCD and filter) was used for all data in the release. This is by virtue that the intrinsic pixel sensitivities can be considered constant to ~0.2% or better for g, r and i (Verdoes Kleijn et al., 2013, ExA, 35, 103). For g, r and i this master flat is a combination of a master dome (for high spatial frequencies) and master twilight (=sky) flat-field (for low spatial frequencies). Both contributing flats are an average of 5 raw flat-field exposures with 3σ-rejection. In u band only the twilight flats are used.
  • Illumination correction. Illumination correction (a.k.a. "photometric superflat") is applied in pixel space, and only on the source fluxes (i.e., after background subtraction). A single illumination correction image is used to correct the single master flats per filter for all KiDS-ESO-DR2 data (see Verdoes Kleijn et al., 2013, ExA 35, 103).
  • De-fringing. De-fringing is only needed for KiDS i band. Analysis of nightly fringe frames showed that the pattern is constant in time. Therefore, a single fringe image was used for all KiDS-ESO-DR2 images observed after 2012-01-11. For each science exposure this fringe image is scaled (after background subtraction of the science exposure and fringe frame) and then subtracted to minimize residual fringes.
  • Pixel masking. Cosmic-rays, hot and cold pixels, saturated pixels are automatically masked as described in MF13 during de-trending. These are included in the weight image. Additional automatic and manual masking is applied on the coadds (see masking page).
  • Satellite track removal. Satellite tracks are detected automatically by applying the Hough transform (Hough, 1962) to a difference image of maximally overlapping exposures within a dither sequence after masking bright stars and bright ghosts. The pixels affected by satellite tracks are masked and included in the weight image.
  • Background subtraction. To remove vignetting by bond wire baffling (Iwert et al. 2006, Proc. SPIE 6276, 62760A) in the focal plane, a row-by-row background subtraction method is used, before the background subtraction done by SWARP (see astrometric calibration page).