This section describes how ommosaic works when the input images are all OM product sky-images.
STAGE 1- Checking of input files
Read in the list of input images from the parameter imagesets Check that all the input files are either sky-images or unrotated images- if not stop the program with an error message. Check that all the files have different names- if not stop the program with an error message
STAGE 2- Some initialization- Inputting of various bits of information from the input files.
Loop over each input file
1) Extract the values for the following FITS keywords in the file header
1) FILTER - the OM filter
2) OBS_ID - the observation identifier
3) RA_PNT - RA pointing direction of OM (degees)
4) DEC_PNT - Dec pointing direction of OM (degrees)
5) POSCOROK - True if astrometry correction applied, FALSE otherwise
6) RAOFFSET - Computed RA offset correction * cos(DEC_PNT) (arcsecs) - 0 if POSCOROK = FALSE
7) DEOFFSET - Computed Dec offset correction (arcsecs)- 0 if POSCORK = FALSE
8) CDELT1 - Plate-scale in RA direction
9) CDELT2 - Plate-scale in Dec direction
10) CRPIX1 - Reference x pixel
11) CRPIX2 - Reference y pixel
12) CRVAL1 - RA value at CRPIX1
13) CRVAL2 - Dec value at CRPIX2
14) BINAX1 - Binning factor (0 or 1)
15) BINBPE - Binning on (True or False)
16) WINDOWDX - Image x-axis length
17) WINDOWDY - Image y-axis length
18) DATE_OBS - Start time of observation
19) DATE_END - End time of observation
20) EXPOSURE - The exposure time (secs)
2) Compute the following quantities
1) RA of bottom left-hand corner of image
2) Dec of bottom left-hand corner of image
3) RA of bottom right-hand corner of image
4) Dec of top left-hand corner of image
5) Modified Julian date of mid-point of exposure
End loop
STAGE 3- Checking the astrometry of the images- Some images may have had an astrometry correction applied to their CRVAL1 and CRVAL2 values. For those that don't, a correction will be applied to CRVAL1 and CRVAL2 by interpolation/extrapolation using a list of the corrections applied to the rest and a list of their mid-point exposure times.
Create an empty list (raList) to store computed astrometry RA offsets
Create an empty list (decList) to store computed astrometry Dec offsets
Create an empty list (timeList) to store the computed mid-observation times
Loop over each input file
If POSCOROK is true then:
1) Add the RA astrometry correction to raList.
2) Add the Dec astrometry correction to decList.
3) Add the mid-point exposure time to timeList.
End if
End Loop
If the size of raList is zero then terminate STAGE 3 and go on to STAGE 4
Sort raList, decList and timeList into increasing time order
Loop over each input file
If POSCOROK is false then:
1) Compute an interpolated/extrapolated RA astrometry correction and replace its RAOFFSET value with it.
2) Compute an interpolated/extrapolated Dec astrometry correction and replace its DEOFFSET value with it.
End if
End Loop
STAGE 4- Adding extra astrometric keywords to the image files FITS headers to say whether or not an interpolated/extrapolated astrometric correction has been computed.
Loop over each input file
If POSCOROK is TRUE then:
1) Add a new boolean FITS keyword POSINTRP to the FITS header and set its value to FALSE.
2) Add a new 32-bit float keyword RA_OFF to the FITS header and set its value to 0.
3) Add a new 32-bit float keyword DEC_OFF to the FITS header and set its value to 0.
else
1) Add a new boolean FITS keyword POSINTRP to the FITS header and set its value to TRUE.
2) Add a new 32-bit float keyword RA_OFF to the FITS header and set its value to the
interpolated/extrapolated value.
3) Add a new 32-bit float keyword DEC_OFF to the FITS header and set its value to the
interpolated/extrapolated value.
end if
End Loop
STAGE 5 - Alignment of images by cross-correlation (This stage is optional and in the omichain will be skipped)
Read in the value of the input parameter mincorr. If its absolute value is less than 0.0001 then skip this stage and go on to STAGE 6. Otherwise, please refer to Section Image Alignment.
STAGE 6 - Creation of the output image, exposure-image and quality-image.
1) Find the maximum left-hand corner RA coordinate (RAL) from all the images
(remember that the RA increase leftwards)
2) Find the minimun right-hand corner RA coordinate (RAR) from all the images
3) Find the minumum bottom edge Dec coordinate (DECL) from all the images
4) Find the maximum top edge Dec coordinate (DECU) from all the images
5) Determine the number of pixels in the x-direction (nx) using the plate-scale
6) If nx exceeds 5000 (possible if using images from different observations),
give a warning message and reset nx to 5000.
7) Determine the number of pixels in the y-direction (ny) using the plate-scale
8) If ny exceeds 5000 (possible if using images from different observations),
give a warning message and reset ny to 5000.
9) Create a 2-d 32-bit float image to store the stacked PRIMARY image.
10) Create a 2-d 32-bit float image to store the computed EXPOSURE image.
11) Create a 2-d 16-bit integer image to store the computed QUALITY image.
12) Set CRPIX1 to nx / 2 and CRPIX2 to ny / 2 (these refer to the output image)
13) Set CRVAL1 to (RAL+RAR) / 2 and CRVAL2 to (DECL+DECU)/2 (these refer to the
output images)
STAGE 7 - Set the pixels in the output PRIMARY, EXPOSURE and QUALITY images.
1) If the images are comprised of a mixture of low/high resolution images set a variable mixedImages to
true- otherwise set it to false.
2) Loop over input image files
Load the image from the PRIMARY block in the file.
If mixedImages=true and the image is a high-resolution one, then multiply each pixel of the image
by four (to effectively convert it into a low-resolution one for interpolation purposes).
Loop over pixels in output image
1) Convert the tangent-plane coordinates, x and y, of the pixel to RA and Dec.
2) Convert RA and Dec to input image tangent-plane coordinates, x1 and y1, using the WCS values
for the image.
3) Check to see that x1, y1 lies within the image, if not move onto the next pixel.
4) Compute the value of the image at x1, y1 by 2-d bilinear interpolation and add this
value to the output image pixel value at x, y.
5) Compute the value of the exposure image at x1, y1 by 2-d bilinear interpolation and add
this value to the output exposure image pixel value at x, y.
6) Compute the value of the quality image at x1, y1 by determining which bits are set at
each of the 9 neighbouring pixels and then by setting the corresponding bits of the quality-image
pixels at x, y.
End loop
End loop
3) Divide each pixel value of the output PRIMARY image by its corresponding EXPOSURE pixel value.
STAGE 8 - Creation of the mosaiced image file.
Determine the name of the output image file from the parameter MOSAICEDSET Create a FITS file with that name.
Add various keywords to the FITS header, including the following-
Add the PRIMARY image to the output file.
Add the EXPOSURE image to the output file.
Add the QUALITY image to the output file.
