4.4 PyDrizzle Products
PyDrizzle combines the separate output images created by drizzle, using the association file, into a single multi-extension FITS file. This file contains a primary header and three image extensions, consisting of a science (SCI), a weight (WHT), and a context (CTX) image.
Example:
In the following example, we present the data products for a simulated 4-point HRC mosaic pattern. Using the HST Phase II proposal language, this pattern is defined as:
- Pattern_Type: ACS-HRC-MOSAIC-BOX
- Pattern_Purpose: MOSAIC
- Number_Of_Points: 4
- Point_Spacing: 24.130
- Coordinate_Frame: POS-TARG
- Pattern_Orient: 90.0
- Center_Pattern: YES
There are two exposures at each pointing (CR-SPLIT=2) to allow for the removal of cosmic-rays, giving a total of 8 raw images. In figure 4.5, we show one set of CR-SPLIT raw images. Note the cosmic rays, hot pixels, bad columns, and overscan regions. Once the 8 raw images are processed and combined by CALACS, the result is 4 calibrated, cosmic-ray rejected images (figure 4.6).
The output from PyDrizzle is a single multi-extension FITS file. The first extension contains the science (SCI) image which is corrected for distortion and which is dither-combined (or mosaiced), if applicable. The drizzled SCI image is presented in figure 4.7 and is in units of counts per second. All pixels have equal area and equal sensitivity across the field of view, giving an image which is both photometrically and astrometrically accurate for both point and extended sources. The dimensions of the output image will be computed on-the-fly by PyDrizzle and the default output plate scale will be read from the IDCTAB. These parameters, however, may be chosen by the user to best suit the actual data. (See table 4.4).
The second extension of the drizzled image contains the weight (WHT) mask. Normal pipeline processing will take into account any bad pixels flagged in the input images' DQ array and will combine these individual weight maps, scaling each by their exposure time. The combined weight mask therefore allows users to verify whether features found in the science image have been affected by bad pixels flagged during calibration.
The weight image for the simulated data is shown at left in figure 4.8 and is scaled to 80% the size of the science image in figure 4.7. Regions which appear "white" are pixels which were flagged in the DQ array of any given input image during processing. These pixels are thus assigned a lower weight than pixels which were not flagged. The majority of the pixels in figure 4.8 are "light grey", indicating that the same number of images were used as input for each of the four pointings, and that these images had equal exposure times. Pixels which appear "dark grey", located in regions where the four images overlap, have twice the exposure time of the "light grey" pixels and thus are given a higher weight.
The third extension of the drizzled image contains a new type of product: the context (CTX) image. This image serves as a bit mask, indicating which input images contributed to each pixel in the drizzled product. Each input image possesses a unique bit value, and the exposure time associated with each bit is stored in the header file of the context image. Thus, with a little work, this array can be used to generate a complete exposure time map for the drizzled science image.
The context image for the simulated data is shown at right in figure 4.8. Because each input image has a unique bit value, the four "quadrants", representing the four pointing positions, have unique "grey" colors. Pixels which are "white" indicate regions which were flagged in both input images. Again, in the overlap regions, the combined bit values are higher than for pixels which have only a single pointing.
Figure 4.5: One of two sets of CR-SPLIT observations, taken using a 4-point dither mosaic with the HRC. These simulated raw images show cosmic rays, hot pixels, bad columns, and the overscan regions. 
Figure 4.6: Calibrated, cosmic-ray rejected images produced by the CALACS pipeline for simulated HRC data. 
Figure 4.7: Science (SCI) extension of the PyDrizzle product for simulated HRC data. This image has been corrected for distortion and drizzled into a single mosaic using the four pointings in the dither pattern. 
Figure 4.8: Weight (WHT) and context (CTX) extensions, left and right, respectively, of the PyDrizzle product for simulated HRC data. These images are displayed at 80% the size of figure 4.7. 
