PSF Photometry for CHEOPS¶
Table of contents¶
Background¶
PIPE is a photometric extraction package for CHEOPS that is complementing the official Data Reduction Pipeline (DRP). While the DRP is using aperture photometry on the D=200 pixel subarrays provided by CHEOPS, PIPE is using PSF photometry on the D=60 pixel imagettes, often provided with a shorter cadence than the subarrays.
Why should you care about PIPE when we already have the DRP? There are a couple of advantages in using PIPE:
The shorter cadence of imagattes increases the time resolution of the photometry. This can be of a distinct advantage for some applications, e.g. when stellar activity such as flares need to be time resolved, or when ingress/egress is desired to be better resolved for timing purposes.
For faint targets PSF extraction makes better use of the signal by weighting it properly over the PSF, so that higher S/N photometry can be extracted.
The shorter cadence and producing an accurate PSF model makes it easier to find and correct for cosmic rays.
PSF photometry can properly deal with hot or bad pixels by either giving them a lower weight or masking them entirely in the fitting process. This may become critical as the detector ages in space, making it increasingly harder to find “good” regions on the detector without bad pixels.
Are there reasons then to not always use PIPE when extracting photometry? Yes, because accurate PSF photometry requires accurate PSF modelling. The PSF in turn depends on
the location on the detector
the spectral energy distribution of the target
the temperature of the CHEOPS telescope tube
the motion blur introduced by pointing jitter
These challenges to produce an accurate PSF are addressed in PIPE but also mean that PIPE is not robust enough to be run blindly, as it can fail if assumptions about the data are wrong. In principle one could attempt to anticipate all ways assumptions can fail, but in practice we are quite far from achieving that. A poorly determined PSF could result in photometric performance that is worse than that provided by the DRP. Thus, PIPE needs to be run in a supervised mode and is not a replacement for the DRP.
Original author: Alexis Brandeker <alexis@astro.su.se>, Stockholm University.