RE: [APML] Gradient and Vignette removal software

["Richard Klappal" <klappal@xnet.com>]

>From "The Image Processing Handbook", 3rd Edition, CRC Press/IEEE Press,
Editor: John C Russ, one of the standard ways to correct for uneven
illumination (should work for both vignetting and sky gradients) is to make
a copy; repeatedly replace each pixel with its darkest neighbor from a 5x5
octagonal neighborhood, until only backgrounds remins; subtract from
original and stretch/relevel the contrast.  Another uses brightest neighbor
from the 5x5 (probably depends on the predominant background).

I suspect that a darkening blur sequence would work similarly, with with
radius dependent on detail size.

Rich


> -----Original Message-----
> From: owner-astro-photo@seds.org [mailto:owner-astro-photo@seds.org]On
> Behalf Of Jerry Lodriguss
> Sent: Thursday, June 27, 2002 02:30
> To: astro-photo@seds.org
> Subject: Re: [APML] Gradient and Vignette removal software
>
>
>
> Hi John,
>
> I NEVER remove stars by hand! <G>
>
> That's what a 100 pixel median filter is for... <G>
>
> Jerry
>
>
> >You are right - this isn't the perfect answer due to film's non-linear
> >response over time, and it won't touch gradients. It might be
> possible that
> >adjusting curves or contrast to the control layer might help offset the
> >effect of reciprocity. Or not, but I do plan to try this. This has to be
> >more accurate and easier that manually removing stars from an image,
> >removing any DSO's, applying a massive blur, inverting and applying as a
> >control layer which is how I do it now. With unfiltered Schmidt shots -
> >which Chris does - exposures are short, generally under 10
> minutes. If all
> >goes well, I will try this out on the weekend.
>
>
> Astronomical photography: http://www.astropix.com
>
>
>
> --  APML Archives at <http://astro.umsystem.edu/apml/>  ---
>              Unsubscribe at <majordomo@seds.org>
>
>



--  APML Archives at <http://astro.umsystem.edu/apml/>  ---
             Unsubscribe at <majordomo@seds.org>