The NICMOS calibration task is divided into two stages: calnica, which is used for every individual exposure, and calnicb, which is used after calnica on those exposures which comprise an association.
The first calibration stage, calnica, ( Figure 12.1) performs those calibrations which can be done to a single exposure using the configuration information from its telemetry and the Calibration Data Base. Such calibrations are derived from the calibration program (see Chapter 13) and typically change on time scales of months. This is analogous to the WFPC2 calibration process (calwp2). Calnica performs the following steps:
MULTIACCUM
data).
MULTIACCUM
data).
Since NICMOS dark current depends on temperature, code has been developed and implemented in the calibration pipeline which generates synthetic, temperature dependent dark reference files for each dataset. More information on temperature dependent darks may be found in the NICMOS Data Handbook.
Figure 12.1: Conceptual calnica Pipeline
Observers will be given both the uncalibrated (raw
) data and the processed data for each exposure. For MULTIACCUM
observations, partially calibrated data for each readout will be generated (which excludes the cosmic ray and saturation corrections), in addition to a final single image.
To recalibrate NICMOS data, observers will need the calnica software (included in the STSDAS
distribution) and the necessary calibration reference files (available from the HST Data Archive using StarView and from the NICMOS WWW pages).
The data processing flow chart for normal imaging and spectroscopic images is shown in Figure 12.2.
Figure 12.2: Calibration Steps of the calnica Pipeline
While previously it has been possible to execute multiple exposures from a single proposal logsheet line (e.g., WFPC2
, CR-SPLIT
, and NEXP
=n constructs), this capability was significantly expanded to support requirements of the second generation and subsequent science instruments. Typical examples include the removal of cosmic rays, the construction of a mosaic image, and the subtraction of the sky background from a sequence of on-target and off-target observations. These observations are distinguished by the fact that their calibration and processing depends upon other observations obtained at the same time.
The calnicb part of the pipeline carries out the calibration and merging of associated data frames, each of which has first been processed by calnica. In the case shown in Figure 12.3, the associated set has 3 individual datasets that are combined into one merged and calibrated dataset.
Figure 12.3: Conceptual calnicb PipelineWe refer to these sets of exposures as associations. The calnicb task operates on an entire association, and produces one or more products from that set ( Figure 12.3). In the case of dither patterns, calnicb reads from the headers of the individual exposure files what the telescope offsets were. It then identifies sources in the images, and, starting with the telescope pointing information from the headers as an initial guess, determines what the pointings actually were. It then combines the images into a final mosaic, rejecting from the output any cosmic rays that had not been detected in the individual exposures when they were processed by calnica. The calnicb code uses all the data quality information generated by calnica to avoid propagating identified cosmic rays, bad pixels, or saturated pixels, into the output mosaic. In the case of chopped images, if multiple images were obtained at each chop position, calnicb generates a mosaic for each background position and produces an output product image for each of those. It then combines each of the background images to determine an average background value for the target position, removes this constant background from the mosaic it has generated for the target position, and then saves the result of this operation as the final, background subtracted mosaiced image of the target. When calnicb is calculating the offsets between images, it starts with the telescope pointing information as its first guess. If it is unable to match the various images by adjusting this pointing information by more than some limit, the code reverts to using the pointing information alone, on the assumption that there are no sources bright enough to detect in the individual images, or that there are instrumental artifacts that are confusing the offset calculation, or that the telescope suffered a loss of guide star lock during one or more of the exposures resulting in a potentially very large offset being introduced. To date, we have found that all of these things happen very rarely.
Space Telescope Science Institute http://www.stsci.edu Voice: (410) 338-1082 help@stsci.edu |