Important Notes: Running this module does not automatically apply the overlap correction. In order to apply the correction and output the corrected BCDs, the Apply Overlap Correction box in this module must be checked.
PURPOSE
This module computes the correction needed to set the background of overlapping images to a constant value. Optionally, it applies the correction to the input BCD images and creates corrected images for input to Mosaic.
INPUT
Top Threshold: (float) Number of sigma above the mean to detect outliers among the computed offsets.
Bottom Threshold: (float) Number of sigma below the mean to detect outliers among the computed offsets.
Minimum Number of Images: (int) Minimum number of overlapping images to perform outlier rejection.
Use Sparse for Many Files: Turn on to use a sparse matrix solver to increase the number of images it can solve.
Skip Bad Image Check: If no bad images, can turn this on for some speed gain.
Apply Overlap Correction: Apply the calculated offsets to the BCD images to create the new background-corrected images.
COMMAND LINE INPUT
&COMPUTEOVRLAPCORRIN
TOP_THRESHOLD = 3,
BOTTOM_THRESHOLD = 3,
MIN_IMG_NUM = 4,
WEIGHT = 0.0,
&END
WEIGHT: (float; command line only) Determines the overall level that the frames should be corrected to. In the first method (WEIGHT = 0.0, the default), it determines the overall level after finding the offsets, by forcing the sum of the offsets of all the images to add up to 0. In the second method (nominally WEIGHT = 1.0), it minimizes the squared offsets as part of the original minimization. See §4.1 Background Matching Overview for more information.
OUTPUT
Overlap Correction Table (offsets.tbl): A table of calculated offsets for each input DCE is the primary output of the Overlap flow.
Corrected Images (overlapCorrection_*.fits): If Apply Overlap Correction is turned on, a new FITS stack is output, with the background-matching offsets applied. These outputs are written to the Overlap_corr subdirectory.
DISCUSSION
The background matching algorithm (see §8.5) calculates an offset to apply to each image in order to set their backgrounds to a constant value. After the initial calculation, the offsets are examined, and outliers are rejected, as long as the minimum number of images are present. The other inputs to this module control how much of an outlier an offset has to be in order to be rejected. A rejected outlier will not be used in determining the overall offset level for all frames, but will still be applied to the corresponding data frame to bring the background to the same offset as the other input frames.
The offsets and their uncertainties are written in the header of the corresponding output FITS files. The keywords are OVRLPDC and OVRLPDCD. They keyword OVRLPOUT is set to 1 if the offset was determined to be an outlier. A table, offset.tbl, is also created, containing the offsets and their uncertainties for all the files. The “outliers” column is used to indicate whether the offset in that row was determined to be an outlier. Below is a sample output table:
