[Aavso-photometry] RE: average vs. median

[Michael Newberry]

Tomas,

That'a a good point about the median---or any outlier rejection method for 
that matter. If you median-combine a group of images taken with even 
slightly different background brightness, the combination returns the image 
of middle brightness, with little or no reduction in noise and little or no 
rejection of any but the most extreme artifacts. You get around that by 
having the software normalize the images as part of the combining process.

Michael Newberry

----- Original Message ----- 
From: "Tom Krajci" <tom_krajci at tularosa.net>
To: <aavso-photometry at mira.aavso.org>
Sent: Friday, January 20, 2006 2:27 PM
Subject: [Aavso-photometry] RE: average vs. median


> >From: "Tomas L. Gomez" <tomas.l.gomez at gmail.com>
>
>>...If, instead of taking the average, I take the median,
>>how is the error reduced?.
>
> Here's one more consideration for median that can get you in
> trouble...changing sky conditions.  If the moon rises or sets during your
> series of images, or thin clouds move in or out, or you image over a long
> enough time that airmass changes...or amount of light pollution 
> changes...a
> median stack of such 'disparate' image sets will not do much good.  If you
> want rejection of way-out pixel values, you may want an algorithm like 
> 'mean
> of the median half'...where the pixels are sorted high-to-low...some 
> highest
> and lowest percentage of this sorted array is discarded (and extreme 
> values
> often are the cosmic ray hits, etc.)...and then the mean is taken of the
> remaining values.
>
> I hope this helps.
>
> Tom Krajci
> Cloudcroft, New Mexico
> http://overton2.tamu.edu/aset/krajci/
>
>
>
> -- 
> No virus found in this outgoing message.
> Checked by AVG Free Edition.
> Version: 7.1.375 / Virus Database: 267.14.21/235 - Release Date: 1/19/2006
>
> _______________________________________________
>
> Aavso-photometry mailing list
> Aavso-photometry at mira.aavso.org
> http://www.aavso.org/mailman/listinfo/aavso-photometry
>
>
>