2 Observations and reduction

M 51 was observed with HST-WFPC2 as part of the HST Supernova INtensive Study (SINS) program (Millard et al. 1999). For this study we use the images taken in the broad band filters F336W (U), F439W (B), F555W (V), F675W (R) and F814W (I) from the SINS program and in the narrow band filters F502N $(\mbox{[OIII]})$ and F656N ( $\mbox{H$\alpha$ }$) from the GO-program of H. C. Ford. The image in U was taken on 1994 May 12, the BVRI images were taken on Jan. 15 1995 and the [OIII] and H$\alpha$ images on Jan. 25 1995. The U and B images were split into three and two exposures of 400 s and 700 s respectively. The [OIII] and H$\alpha$ images are split into two exposures of 1200 s and 500 s ( $\mbox{[OIII]}$), and 1400 s and 400 s ( $\mbox{H$\alpha$ }$). In the remaining bands one single exposure of 600 s was taken. The data was processed through the PODPS (Post Observing Data Processing System) for bias removal, flat fielding and dark frame correction.

To remove the cosmic rays from the U, B [OIII] and H$\alpha$ images, we used the STSDAS task crrej for combining the available exposures. For the VRI images, where only one exposure is available, we used a procedure called "Cosmic Eraser''. This procedure combines the IRAF tasks cosmicrays and imedit to reject as carefully as possible the cosmic rays. The automatic detection of cosmic rays with the task cosmicrays is based upon two parameters, a detection threshold and a flux ratio. The first parameter enables the detection of all the pixels with a value larger than the average value of the surrounding pixels. The flux ratio is defined as the percentage of the average value of the four neighboring pixels (excluding the second brightest pixel) to the flux of the brightest pixel. This parameter allows a classification of the detected objects: cosmic ray or star. Training objects are used to determine the flux ratio carefully. These training objects are labeled by the user to be a cosmic ray or a star. With imedit the detected cosmic rays signals are replaced by an interpolation of a third order surface fit to the surrounding pixels.

After the correction for the cosmic rays, the images were corrected for bad pixels using the hot pixel list from the STScI WFPC2 website in combination with the task warmpix. Corrections for non-optimal charge transfer efficiency on the CCD's of the WFPC2 camera were applied using the formulae by Whitmore & Heyer (1997).

With the task daofind from the DAOPHOT package Stetson (1987), we identified the point sources on the image. We performed aperture photometry on these sources, also with the DAOPHOT package. We used an aperture radius of 3 pixels. The sky background was calculated in an annulus with internal and external radius of 10 and 14 pixels respectively. We only selected the point sources with an uncertainty smaller than 0.2 in the magnitude. Photometric zeropoints were obtained from table 28.2 of the HST Data Handbook (Voit 1997), using the VEGAMAG photometric system (Holtzman et al. 1995).

The aperture correction was measured for a number of isolated, high S/N point sources on each WFPC2-chip. This output was adopted for all the other point sources on the chip. Following Holtzman et al. (1995) we have normalized the aperture correction to 1 $^{\prime\prime}$ (10 WF pixels). The aperture corrections we found are between -0.24 and -0.37 mag. This is larger than the aperture corrections for stars ${\approx}{-}0.17$ mag (Holtzman et al. 1995), which means that the detected point sources are fairly well resolved.

We adopt a distance of $d = 8.4 \pm 0.6$ Mpc (Feldmeier et al. 1997), which corresponds to a distance modulus of 29.62. At this distance, 1 $^{\prime\prime}$ corresponds to a linear distance of 40.7 pc, which means that an HST-WFC pixel of 0.1 $^{\prime\prime}$ corresponds to 4.1 pc.