- Why Drizzle?
- What does the pipeline produce?
- Should I reprocess? How can I improve my drizzled products?
- Reference Material and Examples for WFC3
1. Why Drizzle?
Images produced by both WFC3 channels are affected by considerable geometric distortion, introduced by the tilt of the image surface with respect to the chief ray. This is compounded by non-linear terms that produce changes across the field of view in both plate scale and area subtended by the pixels. Distortion solutions derived from data taken during SMOV with the UVIS F606W filter and the IR F160W filter have been entered into IDCTAB files to support the use of MultiDrizzle to produce distortion-corrected images.
The drizzling process removes the geometric distortion and leaves the sky flat, so photometry of any sources in DRZ images is uniform across the image. This is not true of the FLT images, and therefore a field-dependent correction factor is needed to 1.) achieve uniformity in the measured counts of an object across the field, and 2.) match the output drizzled counts. This correction is called the "pixel area map" and simply reflects the area of the pixels at the location of the source. By multiplying the FLT images by the pixel area map, users will recover the same counts on FLT images and DRZ images.
The MultiDrizzle software was designed to provide a seamless, integrated approach to using the various tasks in the IRAF/STSDAS dither package to correct images for distortion and to register, clean, and optimally combine dithered observations. The algorithm, known as Variable-Pixel Linear Reconstruction (or informally "Drizzle"), was developed by Fruchter & Hook (2002) to combine undersampled dithered images of the Hubble Deep Field and has now been implemented as part of standard processing by the HST calibration pipeline.
Drizzling can be used to combine both dithered and mosaicked exposures. Mosaicking is performed with the aim of increasing the area of sky covered, usually to provide a seamless joining of contiguous frames. Dithering is employed in imaging programs for several reasons, including:
- removal of hot pixels and other detector blemishes
- improving sampling of the PSF
- improving photometric accuracy by averaging over flat-fielding errors
- bridging over the gap between the chips in the UVIS channel
2. What does the pipeline produce?
MultiDrizzle processing was implemented in the WFC3 OPUS pipeline on February 4, 2010. Drizzled data products are produced using the latest on-orbit WFC3 distortion solutions and MultiDrizzle parameter tables, as given in the IDCTAB and MDRIZTAB reference files.
WFC3 Multidrizzle Parameters Tables
WFC3 Distortion Coefficients Tables
To understand the processing which took place in the pipeline, it can be helpful to inspect the MultiDrizzle parameter table or MDRIZTAB. Drizzled products obtained from the archive have been processed using a default set of parameters, as specified in the MDRIZTAB. These parameters work best for observations which were obtained as part of a pre-defined observing pattern and thus are `associated' in the pipeline via an association table (*_asn.fits).
For example, images which were obtained using a sub-pixel dither box pattern are usually aligned to better than 0.1 pixels and have accurate cosmic ray flags. For images obtained in separate visits, the image alignment and cosmic ray flagging must usually be fine-tuned through manual reprocessing. When a sub-pixel dither pattern has been used, the final drizzle sampling can be fine-tuned to produce images with improved overall resolution. The pipeline uses coarse values for both the output pixel size (scale) and drizzling kernel (pixfrac). This speeds up processing of the pipeline and is sufficient to give the user a very good quick view of the field. However, when the user has well dithered data, rerunning MultiDrizzle manually can be useful to derive a more optimal set of parameters.
3. Should I reprocess? How can I improve my drizzled products?
The goal of the WFC3 pipeline is to provide data products calibrated to a level suitable for initial evaluation and analysis for all users. Observers frequently require a detailed understanding of the calibrations applied to their data and the ability to repeat, often with improved products, the calibration process at their home institution. There are several occasions when off-line interactive processing with MultiDrizzle may be desired:
- Combining Images from Multiple Visits
- Optimizing the Image Alignment
- Creating a Large Mosaic
- Improving the Sky Subtraction
- Improving the Cosmic Ray Rejection
- Optimizing the Image Sampling for Dithered Data
Drizzled products from the WFC3 pipeline are created using a default set of parameters which work best when observations are obtained as part of a pre-defined observing pattern. In these cases, MultiDrizzle will produce a high quality combined product from the associated data from a single visit. Future developments to the pipeline will allow MultiDrizzle to identify and combine data from multiple visits of the same target; however, at this time users must manually reprocess their observations to achieve this result. When multiple visits are obtained, guide stars must be reaquired, and the absolute pointing between visits is accurate at the level of the GSC2 (~0.2"). As a result, offsets of a few pixels are typical between visits, and these must be corrected for by the application of a user-define 'shift file'. These additional shifts and/or rotations are used to fine-tune the World Coordinate System (WCS) information to allow for optimal image registration and cosmic-ray rejection.
For targets consisting of large extended or bright sources, the pixel intensity distribution in one or more of the detectors may be significantly skewed toward the bright end by the target itself, thereby overestimating the sky on that detector. MultiDrizzle can be run off-line to limit the range of initial pixel intensity values used in computing the sky and to tighten up the sigma-clipping iterations to produce a more accurate sky subtraction.
Effective resolution can be improved by combining images made with sub-pixel offsets designed to better sample the PSF. This is especially important for WFC3/IR, because the PSF is undersampled by a about factor of 2. When sub-pixel dithered observations are obtained, the user is encouraged to experiment with different combinations of the parameters for the final drizzle combination step: 1.) `scale' (the size of the output pixels) and 2.) `pixfrac' (the linear size of the `drop' in input pixels). One must choose a `pixfrac' value that is small enough to avoid degrading the final image, but large enough that when all images are dropped onto the final frame, the flux coverage of the output frame is fairly uniform.
4. Reference Material and Examples for WFC3
A new MultiDrizzle Handbook has been written to describe the algorithms used by the software, to describe the parameters in detail, and to provide instructions on how to run it, along with an extensive set of examples. This handbook and the accompanying software can be obtained from the MultiDrizzle webpage
To reprocess WFC3 UVIS images, the examples in Chapter 6 developed for the ACS CCDs may be used as a guide (see Section 6.2). Example 1 describes the steps required to improve the drizzle sampling, Example 2 describes how to refine the image alignment (especially important for images obtained in separate visits), and Example 3 describes how to fine-tune the cosmic ray flagging and create a larger image mosaic. To reprocess WFC3 IR images, the example developed for NICMOS described in Section 6.3 may be used as a guide.
WARNINGS: The error arrays of calibrated products produced by CALWF3 (prior to version 2.0) for the IR detector are an underestimate of the true errors. For this reason, users are discouraged from producing final drizzled products which are weighted by the error arrays. Instead, the input images should be weighted according to their individual exposure time.
Created 04/09/2010 MJD