VIII. Post-Cryo Data Release

3. Data Processing

d. Single-Frame Artifact Identification

Contents

i. 2013 Post-Cryo Release

Bright stars produce a number of artifacts in the Post-Cryo single frame images that can trigger spurious source detections or contaminate measurements of real astrophysical sources. These artifacts include diffraction spikes, scattered-light halos, optical ghosts, persistence (latents), and glints. Source detections that are associated with image artifacts are identified by the cc_flags column in the Post-Cryo Single-Exposure Source Database. The methods used to identification and flagging of artifacts during Post-Cryo data processing was based closely upon single-frame artifact flagging algorithms used during processing of the full cryogenic mission data for the All-Sky Data Release, described in section IV.4.g.. In this section we summarize the how the parameters used for artifact flagging were updated to respond to the different characteristics of the Post-Cryo data.

The most significant difference between the artifact flagging in the Post-Cryo data processing and the flagging used in the Cryogenic data processing is the lack of the W3 and W4 bands. Other than the omission of the W3 and W4 artifact flagging, the algorithms used to identify and flag detections affected by artifacts in the Post-Cryo processing were generally the same as used for the full Cryo data processing. The parameters used to identify flag artifacts were retuned to reflect changes in the characteristics of the artifacts in the Post-Cryo data. For each type of artifact, we provide the updated parameters below, along with references to the appropriate Explanatory Supplement section that describes fully the methodology.

Artifact flagging in the second-pass processing of in this Post-Cryo that generated the products for the 2013 Post-Cryo Release improves upon the flagging that was done in the first-pass processing that generated the superseded Preliminary Post-Cryo Release. The improvements addressed primarily how artifacts from glints were tracked, and how adjustements were made in response to different single-exposure background levels.

1. Diffraction Spikes

Diffraction spikes are linear features caused by diffracted light from the telescope's secondary mirror support structure. In a single-exposure image, they extend from a bright source at angles of approximately 45, 135, 225, and 315 degrees, where 0 degrees is aligned with the positive y-axis. A description of flagging methodology for diffraction spikes in single-exposure images can be found in Section IV.4.g.i.1. Below we reprint the functional forms give the updated parameters used to flag diffraction spikes.

The functional form of the equation relating spike length (in arcsec), L_S to parent magnitude, m_p is:

The functional form of the equation relating spike width (in arcsec), W_S to parent magnitude, m_p is:

The values for the parameters are:

For distinguishing between detections that are believed to be spurious and those that are believed to be real but contaminated by a diffraction spike, the function relating Δm_{spur_d} to r_parent is:

To correct for the background level in a single-exposure image, the background corrected diffraction spike length, L_{S_bg}, is related to the original spike length, L_S by the relation:

2. Scattered-Light Halos

Bright sources are surrounded by a scattered-light "halo", essentially the outer wings of an object's point-spread function. This flux can produce spurious source extractions, and can, of course, also result in contaminated photometry for real nearby sources. A description of the scattered light halo flagging methodology can be found in Section IV.4.g.ii.1. Below we reprint the functional forms and parameters used to flag halos.

The function relating the halo radius, r_h (in arcsec), and the parent magnitude, m_p

The values of the parameters are:

The function relating the determination of spurious vs. contaminated sources is given by:

Where 'a' and 'c' are the derived parameters. The values of the parameters are:

The size of the halos scales with the background level in the image by the relation:

where B is defined as:

The values c_bg and d_bg are:

3. Optical Ghosts

Optical ghosts are the produced by internal reflections in the optical path of the telescope and camera. In the WISE images, ghosts manifest themselves as ring-like structures at a fixed position from a bright parent source. Since, for a given band, ghosts always manifest themselves in the same position (direction and distance) relative to the center of a bright parent star, flagging is performed using a purely positional approach. Sources are flagged inside a circular region located at a fixed distance and direction from the parent star's center. Although the ghosts themselves are not circular, a circular flagging region is a good first order approximation. The size of a ghost does not vary with the brightness of the parent star.

A description of the flagging methodology for optical ghosts can be found in Section IV.4.g.iv.1. The parameters used to flag ghosts in Post-Cryo second-pass processing are summarized below.

Table 8 contain the parameters used for flagging optical ghosts, including: Δx and Δy, which define the positional offset (in pixels) of the ghost from the parent source; R_ghost, which defines the ghost size; m_{thr_o}, which is the parent brightness in each band at which ghosts appear; and Δm_{spur_o}, the threshold for real vs. spurious detections.

4. Latents (Persistence)

The following section contains information regarding the parameters used persistence (latent) artifact flagging during Post-Cryo second-pass processing. See Section IV.4.g.iv.1 for the characterization methodology and definition of the parameters.

5. Glints

Glints are artifacts on single-exposure images that occur when a bright source just off-frame produces reflected light onto the detector. A description of glint types and their flagging methodology can be found in Section IV.4.g.v.1. Below we give the parameters used to flag glints in second-pass Post-Cryo processing.

6. Cautionary Notes

1. Artifact flagging in the Post-Cryo data processing was not perfect. Parameters were tuned to provide good performance on average for a large range of data conditions, but could not be made to operated ideally for all WISE single-exposure image data. The best way to determine if a source may be contaminated by or a spurious detection of an image artifact is to examine it by eye on the images.


2. For the cleanest possible sample of sources (with respect to artifacts), we requiring cc_flags = '0000' in your database searches. However, as noted above, there are unflagged artifacts, so examine the images.


3. In the galactic plane, artifacts are often overflagged because of the complex and high background level in all bands. The background scaling procedure partially mitigates this effect, but does not do a perfect job for the highest background levels. You may wish to allow contaminated (lower-case) flags for all artifacts if your search is within a degree or two of the galactic plane.


4. Glints appearing on the cross-scan sides of single-exposure images are not flagged.


5. As noted above, glint flags are always lower-case in cc_flags. If your source has a 'g' flag value in a band, inspect the source on the image to make sure it is not a spurious extraction.


6. In rare cases, a source may be bright enough that it produces a ghost which has a latent. Latents from ghost images are not flagged.

ii. Post-Cryo Preliminary Release

The Post-Cryo Preliminary Data Release was generated using first-pass processing of the WISE Post-Cryo survey data. This processing used parameters that were adapted from the Cryo survey data processing and were not necessarily optimized for the characteristics of the Post-Cryo data. Post-Cryo Preliminary Release Single-Exposure Source Database entries that may be spurious extraction of or contaminated by artifacts produced by bright stars are identified by the cc_flags database column. Flagging of sources affect by artifacts in Post-Cryo first-pass processing was based closely on the method used for the first-pass processing of the Cryogenic data, and is described in section IV.3.f. of the Preliminary Data Release Explanatory Supplement. In this section we summarize important characteristics which distinguish flagging in the Preliminary Post-Cryo Release from subsequent processing used to produce the WISE All-Sky Data Release, the 3-Band Cryo Data Release and the 2013 Post Cryo Release.

Some things to keep in mind about artifact flagging in the Post-Cryo Preliminary Data Release:

1. The first-pass processing version of ArtID does not include adjustments for background level of each frame. Spike lengths as well as latent and halo sizes may be overestimated near the galactic plane. Differences in parent magnitude thresholds in areas of locally high background may result in overflagging for all artifacts.


2. The first-pass version of ArtID does not flag glints. It is strongly recommended that you visually examine sources on the Single-exposure images to ensure that none are the result of glints from off-frame bright stars.


3. Refer to section IV.3.f. of the Explanatory Supplement to the WISE Preliminary Data Release for a description of all artifacts and parameters.

The cautionary notes from the preliminary processing are also applicable to post-cryo preminimary processing. They are listed again here for your convenience. Read them!

1. Diffraction Spikes and Scattered-light Halos

In preliminary processing, diffraction spikes and halos are treated together as a single artifact. Please read section IV.3.f.i. of the Explanatory Supplement to the WISE Preliminary Data Release for a description.

Cautionary notes:

• Since Δm_{spur_d} is a fixed constant, it will not be perfect in determining whether an object is real or spurious. The threshold is tuned to favor reliability over completeness. That is, Δm_{spur_d} is intentionally tuned to a value that is smaller than the "ideal" value in order to ensure that no spurious detections are flagged as real sources. Thus, there may be a significant number of real sources near bright stars that are flagged a spurious.


• Values of L_S, W_S, and R_halo for diffraction spikes are dependent only on band and parent magnitude and do not incorporate other factors which may affect spike parameters (e.g., background level). Thus sources may be overflagged in areas of locally high background, where spikes tend to be shorter and halo radii smaller.

2. Optical Ghosts

Please read section IV.3.f.ii. of the Explanatory Supplement to the WISE Preliminary Data Release for a description.

3. Latents (Persistence)

Please read section IV.3.f.iii. of the Explanatory Supplement to the WISE Preliminary Data Release for a description.

Cautionary Note:

• There is a fixed flagging radius for all types of latents in the WISE Preliminary Data Release single frame source catalog. The 8 pixel radius is larger than the size of most short-term latents for in W2. Since the long-term latents persist for up to 12 hours, many more sources may be flagged for long-term latents than are substantially affected in the preliminary post-cryo release. The best way to confirm that a source is affected by a latent is to look at the image.

4. Glints

Glints are not flagged in first-pass Post-Cryo processing.

Last Update: 2013 April 18