wiki:BuildSevenWork

CALCORON3 Development for Build 7

The current outline for CALCORON3 processing as developed by the JWST Calibration Working Group is here. It consists of the following steps:

  1. Assemble Reference PSF's
  2. Align Reference PSF's
  3. Reference PSF Subtraction
  4. Outlier Detection
  5. Image Combination

The JCCWG list of steps also contains an item Create Exposure Level Products, which for coronagraphic mode processing consists only of updating the input exposure-based products to contain DQ flags resulting from the Outlier Detection step. Adding the DQ flags to the input products will be done as part of the Outlier Detection step itself and hence a separate CALCORON3 step is not needed for this.

Step Details

Assemble Reference PSF's

This step will need to assemble all integrations (not exposures, but rather all individual integrations from each exposure) from input exposures designated in the ASN table as being reference PSF. No data/exposures/integrations of the science target will be included, at least not for the Build 7 baseline pipeline. The reference PSF integrations need to be collected and rearranged into a form that is suitable for input to the next CALCORON3 pipeline step, which is alignment of the PSF integrations with the target integrations. More than one reference PSF observation can be specified, which implies multiple input PSF exposures, each of which will be in the form of a _calints product, because we need to use the individual integrations throughout CALCORON3 processing. All integrations from all input PSF exposures should be assembled into a single new, intermediate product, which will be a larger 3D cube model containing all integrations from all PSF exposures stacked into a single cube.

Input(s): List of _calints (3D) products from ASN table labeled as being reference PSFs
Output: Single product (3D) containing stack of all PSF integrations from all reference PSF exposures

Align Reference PSF's with Target

The image alignment code provided by the Coronagraphic Working Group (CWG) is currently set up to work with a single 3D cube of images and aligns each image plane in the cube relative to a single designated plane in the cube (e.g. align all image planes relative to plane zero). This process needs to be repeated for each integration of the input science target exposures, resulting in a new aligned stack of PSF images for each target integration. The flow of the step will be:

    Loop over science target exposures (as specified in ASN table):
        Loop over integrations in each target exposure:
            Align all integrations in PSF product stack to target integration
            Append aligned PSF integrations to target output product (4D stack)

An intermediate 4D product will be created for each target exposure, where the 4D array contains the aligned PSF integrations corresponding to each target integration. The axes of the 4D array will correspond to target_integration x PSF integration x rows x cols.

Inputs: Stacked PSF product from step 1; One or more science target exposures from ASN table, in the form of _calints products.
Output(s): One 4D product per input science target exposure containing aligned PSF images.

PSF Subtraction

PSF subtraction is accomplished using the KLIP algorithm, which first creates a best-fit PSF for the target image and then subtracts the fitted PSF from the target image. This operation will be performed for each integration of each science target exposure, using the aligned PSF images from step 2. The inputs will be the same set of 3D (_calints) science target exposures used as input to step 2, as well as the set of 4D aligned PSF products produced by step 2 (one for each science target exposure). The outputs will be PSF-subtracted versions of each input science target exposure, as well as products containing the fitted PSF's for each target exposure.

Inputs: One or more science target exposures from ASN table, in the form of _calints products. One or more 4D aligned PSF products from step 2 (one per science target product).
Ouputs: One or more PSF-subtracted science target exposures. One or more fitted PSF products (one per science target product).

Outlier Detection

This step will be used to identify and flag outliers (CR hits) that did not get removed during the jump and ramp_fit steps of calwebb_sloper processing. Coronagraphic exposures will often contain relatively few groups per integration, making ramp-based CR detection difficult. The algorithm will be essentially the same as that used in calwebb_image3 for normal imaging mode observations. The only difference here is that the detection scheme should be applied to a single exposure at a time, working only on the stack of integrations within that exposure. Because alignment is expected to be quite good throughout the course of a single exposure, the detection can be done in the detector space of each exposure, with no need for first resampling the integrations to a common output frame. The inputs will be the individual PSF-subtracted science target exposures produced in step 3. Detected outliers will be flagged in the DQ arrays of both the original input science target exposures (listed in the ASN table), as well as the PSF-subtracted target products that are being analyzed.

Inputs: One or more PSF-subtracted science target exposures (from step 3).
Outputs: Updated versions of the PSF-subtracted science target exposures.

Image Combination

This step will take all of the integrations from all of the PSF-subtracted and CR-flagged science target exposures produced by the previous steps and combine them using a drizzle technique, as is done for regular imaging modes. The only difference here is that each integration in each input exposure will need to be treated as a separate input image to the drizzle process. This assumes that the WCS information in the input exposures will faithfully reflect any differences in roll angle and absolute pointing, so as to properly align all of the inputs when combining them.

Inputs: One or more PSF-subtracted and CR-flagged science target exposures (from step 4).
Outputs: A single combined science target image.

Last modified 2 years ago Last modified on 08/03/16 11:37:36