Return to Previous Section

E. Strategy for Obtaining Dark Frames


The strategy for obtaining dark frames to be used to calibrate actual sky data should be somewhat different from the strategy outlined above to test the dark current of our detector. When taking darks one must always keep in mind the trade-off between the integration time devoted to darks (that cannot be used to observe the sky) and the signal to noise ratio of the final dark image. The following set of guidelines should allow observers using our system to obtain dark frames adequate for most purposes:

The reason for this last point has to do with correction for the presence of cosmic rays. If one takes three frames and medians them, one will very effectively remove the cosmic rays in most cases. Obviously, the larger the number of dark frames, the better. This is true both for the removal of cosmic rays and for the improvement of the overall signal to noise ratio. As an example of this, Figure 15 shows two dark frames. Both are from the same night. The one on top is a 30-minute bias-corrected dark. The one on the bottom is a medianed combination of the 10-, 20- and 30-minute darks taken at that temperature. This last is a bit of a cheat, as the darks were not all of the same exposure time. Consider, however, that one would need to devote 1.5 hours of observing time to obtain three 30 minute darks. There are two points to notice from Fig. 15. First, the bottom image is smoother overall. That is, it has higher pixel-to-pixel signal to noise ratio as a result of having twice the exposure time of the top image. Second, the cosmic rays that are visible in the top image have been effectively removed by the medianing process, as advertised.

  • Figure 15: Top: A 30-minute dark frame with T ~ -35 degrees C, obtained on 18 January 2004. Bottom: A scaled median of three darks with T ~ -35 degrees C, all obtained on 18 January 2004.


    Procede to Next Section

    Return to Main Dark Page

    Return to Outline


    Updated: 2009 August 20 [pbe]