2 Observations and reduction of the frames

Observations

The broad band B and R frames were taken during three observing runs in February 1995, May 1996 and June 1997 with the CAFOS focal reducer CCD camera attached to the Cassegrain focus of the Calar-Alto 2.2m telescope. Some details of the observations are given in Table 1.

 

 

Table 1: Dates and set-up of direct imaging with the Calar Alto 2.2m telescope and CAFOS

Dates	Detector	Detector	Pixel	Pixel	Photm.	Exposure	Sky	Seeing
	type	size	size	scale	band	time	bright.	FWHM
			[$\mu$m]	[$''\!\!$/px]		[s]	[ $\frac{\rm mag}{\ifmmode\hbox{\rlap{$\sqcap$ }$\sqcup$ }\else{\unskip\nobreak\hf... ...rlap{$\sqcap$ }$\sqcup$ } \parfillskip=0pt\finalhyphendemerits=0\endgraf}\fi''}$]	['']
1995, Feb. 02-03	Tek#13	1024$\times$1024	24	0.489	B	60, 600	22.4	1.8
					R	60	20.9	2.1
1996, May 20-24	Loral 8	1024$\times$1024	15	0.333	B	60, 600	22.3	1.7
1997, June 12-15	Site#1d	2048$\times$2048	24	0.531	B	300, 600, 900	22.0	1.5
					R	150, 300, 600	20.9	1.5

Most of the data were either obtained for aquisition purpose of the following spectroscopy or as a snap-shot survey to obtain total magnitudes. Typical exposure time for every observing run is printed bold in Table 1. Dedicated deeper (900 s B and 600 s R) images were observed for all isolated ELGs. A set of bias and flat frames for CCD corrections was taken, as usual. For magnitude calibrations and colour corrections we observed during each night a number of fields with standard stars in the star clusters M 92, NGC 2264, NGC 2419, NGC 4147 and NGC 7790 (Christian et al. [1985]). During the May 1996 campaign the spectrophotometric standard stars HZ 21, HZ 44, and BD +33#2642 were observed and used for photometric calibrations.

Zeropoint magnitudes (c _0,i) and colour coefficients (c_1,i) of the form

$$B~=~b+c_{0,B}+c_{ 1,B}(b-r)-c_{ 2,B}X, \protect$$ (1)

$$R~=~r+c_{ 0,R}+c_{1,R}(b-r)-c_{ 2,R}X, \protect$$ (2)

were fitted to the instrumental data of every observing night, where B, R are standard magnitudes, b, r are the instrumental magnitudes and X is the airmass. We found a significant colour coefficient c_1,B during all three observing runs and applied it to the magnitude calibration. We used the following mean values of the extinction coefficients: c_2,B= 0.2 and c_2,R = 0.1.

Data reduction and filtering

The raw CCD frames were de-biased and flat-field-corrected, as described by Stickel et al. ([1993]). Further data reduction was done by means of the Potsdam Image Processing System (PIPS) running within MIDAS environment. The bias and flat-field corrected CCD frames were searched for cosmic ray hits and possible hot and cold pixels by looking for pixel values above the expected noise and checking if they had a point spread function smaller than the estimated seeing. By means of Laplace filtering such faulty pixels were detected, masked out and replaced by the mean value of the surrounding area using the background interpolation routine of the PIPS. To improve the signal-to-noise ratio, we applied the adaptive filtering technique described in Lorenz et al. ([1993]). The main advantage of this technique consists in recognition of the local signal resolution and adapting its own impulse response to this resolution. The maximum filter size and the strength of the filter are variable. We varied the maximum filter size between 15$\times$15 pixels and 23$\times$23 pixels depending on the quality of the frames. The filter strength, defined by the minimum signal to noise ratio for the detection of a local signal, was generally between 2.0 - 2.5 $\sigma$, where $\sigma$ is the rms noise level of the sky background at each scale length.

After the filtering, a careful sky background level determination and subtraction was performed on the smoothed frame, as described in more detail in Vennik et al. ([1996]). After the background subtraction the galaxy image was extracted from the large CCD frame and interactively cleaned from disturbing objects like e.g. foreground stars projected onto the galaxy. We applied an interactive polygon editor for this cleaning procedure.