Data reduction

All image reduction and data analysis was performed in the IRAF⁵ environment⁶.
From previous runs at Calar Alto there was a necessity to use dark frames, although the dark correction should have been in principle included by doing a sky subtraction with a sky frame. Tests showed that the OMEGA-PRIME camera images did not require dark subtraction as we saw no differences between the resulting images.
Creating flat field frames was the next step. Tests on our data made us decide to use sky flat-fielding frames, because those provide flatter final images compared to dome or twilight images.
To reduce the noise and remove background sources it is preferred to use as many sky frames as possible to form a mastersky for the subtraction of the sky frames from the object image, but this is limited by the rapid changes in the sky structure. However, as tests showed we were limited by this, we had to use the minimal number of sky frames for the object images. So to remove the influence of the stars, SExtractor was used to mask stars on the OFF images, creating an output mask image for each sky frame. Of all objects in the sky image, stars and galaxies, the shape of the flux area is determined by SExtractor and a mask area was created, resulting in an output SExmask image belonging to each object image. Interpolation of the masks with the background value did not give good results as we still saw residuals in the final mosaic-ed image. Because we had many images to be stacked, we decided to keep track of the masks and completely remove possible residuals, and thus we did not need to interpolate. The separate SExmask images were used in the final combination to mask the holes produced by stars in the sky image.
Before every object frame was subtracted with the appropriate sky frames (the sky frames that were taken closest to them in time), in this case the minimum number of sky frames (two) near the object frame were used for all reduced galaxies. The sky subtracted target frames were then divided by the sky flat-field (the combination of the two sky images which were normalised). Each of these reduced frames were inspected in case they had large gradient patterns that is not removed by sky subtraction in the combination. Those frames were thus rejected for the final image combining.
The next part consisted of defining relative spatial offsets betweens each object frame in the data (mosaic-ing). This was done by taking three clear and bright stars around the galaxy, ideally forming a triangle, and marking their position in each object frame. After the offsets were determined, all images were combined using a list of the masks for each image that had been made previously with SExtractor so that all the residual influence of the stars in the sky images were removed. Typically a handpicked statistics section (to additively scale the background of each object image to each other) on an already flat part in the to be combined images was selected for the combination - so that the residual background was scaled to an already very flat region.

TABLE 3

MEDIAN OR MEAN COUNTS IN SMALL BOXES

Galaxy	Filter	Mean of Median	Mean of Mean	Median of $\sigma^c$
		[counts]	[counts]	[counts]
(1)	(2)	(3)	(4)	(5)
IC 3322A	J	4.32 $\pm$ 0.25	4.38 $\pm$ 0.25	1.10
IC 3322A	K'	-16.81 $\pm$ 0.13	-16.77 $\pm$ 0.13	1.89
NGC 2424	J	-0.61 $\pm$ 0.23	-0.46 $\pm$ 0.23	1.75
NGC 2591	K'	-11.24 $\pm$ 0.17	-11.31 $\pm$ 0.17	1.27
NGC 5290	J	53.00 $\pm$ 0.21	53.27 $\pm$ 0.21	1.34
NGC 5290	K'	16.56 $\pm$ 0.11	16.68 $\pm$ 0.11	1.40
NGC 5348	J	-38.51 $\pm$ 0.23	-38.09 $\pm$ 0.23	1.76
NGC 5981	J	18.76 $\pm$ 0.31	18.73 $\pm$ 0.31	1.18
IC 2531	J	-0.84 $\pm$ 0.20	-0.83 $\pm$ 0.20	1.79
IC 2531	K'	-2.79 $\pm$ 0.28	-2.66 $\pm$ 0.28	3.17
NGC 0973	J	-0.32 $\pm$ 0.16	-0.33 $\pm$ 0.16	1.52
NGC 0973	K'	-2.31 $\pm$ 0.33	-2.40 $\pm$ 0.33	3.28
NGC 1886	J	-0.12 $\pm$ 0.32	-0.19 $\pm$ 0.32	1.76
NGC 1886	K'	-0.62 $\pm$ 0.44	-0.65 $\pm$ 0.44	3.74
NGC 2424	J	-0.25 $\pm$ 0.46	-0.40 $\pm$ 0.46	1.45
NGC 2424	K'	-1.55 $\pm$ 0.72	-1.59 $\pm$ 0.72	2.45
UGC 3186	J	0.28 $\pm$ 0.18	0.26 $\pm$ 0.18	1.33
UGC 3186	K'	-0.38 $\pm$ 0.29	-0.43 $\pm$ 0.29	2.65
UGC 4277	J	-0.68 $\pm$ 0.16	-0.68 $\pm$ 0.16	1.50
UGC 4277	K'	-0.43 $\pm$ 0.46	-0.51 $\pm$ 0.46	2.74
FGC 2339	R	3888.66 $\pm$ 5.46	3887.30 $\pm$ 5.46	24.23
IC 5249	R	4564.68 $\pm$ 14.45	4569.42 $\pm$ 14.45	26.56
NGC 4179	V	563.6 $\pm$ 1.72	564.2 $\pm$ 1.72	4.15
Notes: (1) Galaxy. (2) Filter. (3) Mean pixel value of the median values of the small boxes made on the sky background and the standard deviation to this mean. (4) Mean pixel value of the mean values of the small boxes made on the sky background and the standard deviation to this mean. (5) Median of the standard deviation in the small boxes made on the sky background.

Ideally the final background of the image is zero. However, this is rarely the case. Thus the value of the background has to be subtracted from the image. The background however, is not always perfectly smooth and flat. There are residual large scale gradients and of course forground stars and background galaxies. To find the remaining background value sections around the galaxy were made in the shape of small boxes, 20$\times$20 pixels in size. They were placed manually around the galaxy to map the background as good as possible avoiding the influence of starlight. The mean of the median value of all those boxes provided our background pixel value. The results are listed in Table 3, together with the standard deviations of those means and for comparison the mean value of the boxes when taking the mean value. After the substraction of this value the image was ready.