Dangers of cosmicrays task?
Sean wrote on Jul 30, 2007
Hi,
I am reducing about 500 images taken with the Johns Hopkins Adaptive Optics Coronagraph on the Palomar 60 telescope. Most of the images have between a 500 and 1000 second integration time, and, thus, have quite a lot of cosmic ray hits. In order to remove all of the hits I think I will need to clean very aggressively (ie. flux ratio ~ 15%, threshold = 3 - 4 sigma). I am curious of what exactly the dangers of such aggressive cleaning are, and how the data might be affected. Due to the characteristics of the instrument, the stars have a FWHM of about 7, so I would not think that cosmic ray cleaning would be likely to affect them. Also, it seems to me that, if the cleaning clips the top of the noise distribution, it would do so evenly, and therefore not affect the photometry. Could you please tell me if there are some nuances of the task (or photometry in general) that I am missing?
~Sean
I am reducing about 500 images taken with the Johns Hopkins Adaptive Optics Coronagraph on the Palomar 60 telescope. Most of the images have between a 500 and 1000 second integration time, and, thus, have quite a lot of cosmic ray hits. In order to remove all of the hits I think I will need to clean very aggressively (ie. flux ratio ~ 15%, threshold = 3 - 4 sigma). I am curious of what exactly the dangers of such aggressive cleaning are, and how the data might be affected. Due to the characteristics of the instrument, the stars have a FWHM of about 7, so I would not think that cosmic ray cleaning would be likely to affect them. Also, it seems to me that, if the cleaning clips the top of the noise distribution, it would do so evenly, and therefore not affect the photometry. Could you please tell me if there are some nuances of the task (or photometry in general) that I am missing?
~Sean
Francisco Valdes wrote on Jul 30, 2007
Hello Sean,
You have correctly stated that the main scientific danger is clipping of the cores of stars. Even a FWHM of 7 pixels might result in the central pixel of a bright star appearing above the threshold relative to the surrounding pixels. The effect of clipping noise spikes in the sky is not a significant concern as you also suggest. Note that most cosmic ray cleaning algorithms do reasonably well on single pixel events but not so well on multiple pixel events. It is generally more powerful if you have "split" observations or other way to compare exposures but I know this is sometimes not possible.
I hope this helps.
Yours,
Frank Valdes
You have correctly stated that the main scientific danger is clipping of the cores of stars. Even a FWHM of 7 pixels might result in the central pixel of a bright star appearing above the threshold relative to the surrounding pixels. The effect of clipping noise spikes in the sky is not a significant concern as you also suggest. Note that most cosmic ray cleaning algorithms do reasonably well on single pixel events but not so well on multiple pixel events. It is generally more powerful if you have "split" observations or other way to compare exposures but I know this is sometimes not possible.
I hope this helps.
Yours,
Frank Valdes
Last post on Jul 30, 2007