Re: [APML] OT: Deep M81 (12 Hours) (My image...really!)

[Roland Roberts <roland@astrofoto.org>]

-----BEGIN PGP SIGNED MESSAGE-----

>>>>> "rg" == Robert Gendler <robgendler@worldnet.att.net> writes:

    rg> Deconvolution algorthms are complex mathematical formulas that
    rg> try to approximate what the image would look like without the
    rg> blurring effects brought on by seeing, mechanical and optical
    rg> issues.

They're not *that* complex :-)

What most folks may not realize is that from a mathematical perspective
you've been doing convolutions for a long time, every time you take a
picture.  The optics of you lens/scope, atmospheric seeing, etc., all
serve to take those perfect point-like sources and spread them around.
For a discrete image (CCD or scanned film), imagine doing this by taking
a perfect image and applying a blurring filter, say a Gaussian (since
that's what the result of all those noise sources ends up becoming).  A
perfect one-pixel star suddenly bleeds into adjacent pixels.  So does
every other light source.

A deconvolution filter is one that is the mathematical inverse of that
Gaussian blur.  It's more than a sharpening filter in that it attempts
to be a true mathematical inverse.  The problem is that you can't tell
from the final image what the correct inverse is.  Good software
packages can guess, and some can even calculate an inverse based on you
selecting a "typical" star and telling it this is your point spread
function (PSF); the PSF is the convolution function, so the software
will invert this.  I believe IRIS has a maximum entropy deconvolution
function which does this using the supplied PSF (selectable from a
stellar image in the field).

In fact, a convolution filter and a deconvolution filter work exactly
the same way mathematically, it just that they happen to "multiply" to
give you no net effect on your image.

regards,

roland
- -- 
                       PGP Key ID: 66 BC 3B CD
Roland B. Roberts, PhD                             RL Enterprises
roland@rlenter.com                            6818 Madeline Court
roland@astrofoto.org                           Brooklyn, NY 11220

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3ia
Charset: noconv
Comment: Processed by Mailcrypt 3.5.8, an Emacs/PGP interface

iQCVAwUBQjXyQeoW38lmvDvNAQE6gAP+OooSJDwCgGFNhYB6Yb9aWiDmle0/BJ3H
fWLUDxv/oIaHiPB/OcblvYkZFCIXFiQemhIAAnvliCZYKO8Un/ZUeDhr15PY+yIh
yy+10mZE3ug6Rpy1qE0ogwgsiUjAJztwQQjQPzRIexAGiYBaq0H0hwihU9KC9o7U
UlkyOkaaQ+k=
=alsm
-----END PGP SIGNATURE-----
_______________________________________________
Astro-Photo mailing list
Astro-Photo@seds.org
http://seds.org/mailman/listinfo/astro-photo