Re: Focus indication

[Tom Droege <tdroege@veriomail.com>]

Norman Molhant (are you still here?) proposed this method long ago.  I 
think it would work with clear sky.  The multi image on one frame has the 
advantage that it works even with clouds.  But it is not so automatic.

I don't try to solve all the problems, but I do try to formulate some that 
are a pain for me.  I just hope that those of you with more software skills 
than I have will go after one or more of these problems.

Tom Droege

At 09:15 AM 8/6/01 -0700, you wrote:

>If you want a focus indication why not choose a method that is
>applicable to all images, not just images of the night sky?
>Auto focus SLR cameras don't detect stars or compute the FWHM
>of anything.
>
>An in-focus image has higher spatial frequency content then an
>out of focus image.  Or equivalently, an out of focus image is
>like a focused image which has had a low pass filter applied to
>it.
>
>One thing to do is to compute the mean delta between adjacent
>pixels.  Every non-edge pixel (All four million of them) has
>four adjacent pixels.  Take an average of all 16 million
>differences or maybe just use the central 500 x 500 pixels.
>This should allow you to focus on anything you point the camera
>at, the sky or even your dog.
>
>
>--
>    Chris Albertson
>    chrisalbertson90278@yahoo.com
>    Redondo Beach, California
>    home: 310-376-1029
>    cell: 310-990-7550