Re: Can out of focused pics be repaired once in the PC?

Discussion in 'Digital Photography' started by josvanr@xs4all.nl, Dec 9, 2004.

  1. Guest

    Isn't it possible to somehow measure the distortion for a particular
    lense, using a known calibration pattern that is photographed and then
    apply a system-theory approach, i.e. treat the camera as a black box?
    Jos
     
    , Dec 9, 2004
    #1
    1. Advertising

  2. Mark² Guest

    <> wrote in message
    news:...
    >
    > Isn't it possible to somehow measure the distortion for a particular
    > lense, using a known calibration pattern that is photographed and then
    > apply a system-theory approach, i.e. treat the camera as a black box?
    > Jos


    Only in the movies.
    :)
     
    Mark², Dec 9, 2004
    #2
    1. Advertising

  3. Mark² Guest

    "Mark²" <mjmorgan(lowest even number here)@cox..net> wrote in message
    news:6FUtd.422871$a85.304648@fed1read04...
    >
    > <> wrote in message
    > news:...
    > >
    > > Isn't it possible to somehow measure the distortion for a particular
    > > lense, using a known calibration pattern that is photographed and then
    > > apply a system-theory approach, i.e. treat the camera as a black box?
    > > Jos

    >
    > Only in the movies.
    > :)
    >


    Unless you're talking about distortion ONLY...this CAN be done, and focus
    can be recovered to a very small degree.
    (I read the subject line, which asked only about focus)
     
    Mark², Dec 9, 2004
    #3
  4. Martin Brown Guest

    wrote:

    > Isn't it possible to somehow measure the distortion for a particular
    > lense, using a known calibration pattern that is photographed and then
    > apply a system-theory approach, i.e. treat the camera as a black box?
    > Jos


    It depends how out of focus they are and how much you are prepared to
    pay to have it done. Deconvolution codes do exist and can work well
    enough to be useful on some problems - motion blur for instance.

    Usually it is much easier to take the shots again.

    Regards,
    Martin Brown
     
    Martin Brown, Dec 9, 2004
    #4
  5. RustY© Guest

    <> wrote in message
    news:...
    > Isn't it possible to ........treat the camera as a black box?
    > Jos
    >


    No! You only have the information from the distorted picture.
    --
    For Welsh Military Flying visit .......
    www.groups.yahoo.com/group/V-A-S/
     
    RustY©, Dec 9, 2004
    #5
  6. Ryadia Guest

    <> wrote in message
    news:...
    >
    > Isn't it possible to somehow measure the distortion for a particular
    > lense, using a known calibration pattern that is photographed and then
    > apply a system-theory approach, i.e. treat the camera as a black box?
    > Jos
    >

    There exists a program called "unshake" which will remove the 'other' images
    from an image. Another is called "focus Fixer". Much is claimed for focus
    fixer but my experience tends to show it is nearly useless. Unshake on the
    other hand is only of use when the photo is sharply focused but you moved
    the camera. This program actually does work but it's really odd interface
    will turn a lot of people off.

    Doug
     
    Ryadia, Dec 9, 2004
    #6
  7. Ron Hunter Guest

    wrote:
    > Isn't it possible to somehow measure the distortion for a particular
    > lense, using a known calibration pattern that is photographed and then
    > apply a system-theory approach, i.e. treat the camera as a black box?
    > Jos
    >


    There are programs that attempt to correct for focus and movement
    problems. I have found them slow, and rarely do they produce
    satisfactory results. The best that can be said for any I have tried,
    is that they make the effect less visible. There is no substitute for
    doing it right in the first place.
     
    Ron Hunter, Dec 9, 2004
    #7
  8. P.R.Brady Guest

    wrote:
    > Isn't it possible to somehow measure the distortion for a particular
    > lense, using a known calibration pattern that is photographed and then
    > apply a system-theory approach, i.e. treat the camera as a black box?
    > Jos
    >

    I understood that some techniques like this were used for early Hubble
    images, but they still had an expensive trip to fit a correcting lens!

    Phil
     
    P.R.Brady, Dec 9, 2004
    #8
  9. Mark² Guest

    "P.R.Brady" <> wrote in message
    news:...
    > wrote:
    > > Isn't it possible to somehow measure the distortion for a particular
    > > lense, using a known calibration pattern that is photographed and then
    > > apply a system-theory approach, i.e. treat the camera as a black box?
    > > Jos
    > >

    > I understood that some techniques like this were used for early Hubble
    > images, but they still had an expensive trip to fit a correcting lens!


    Which indicates that it doesn't work well...
     
    Mark², Dec 9, 2004
    #9
  10. Bill The Cat Guest

    This is not true, or at least there is an exception. Several blurred images of the same subject taken from the same position at the same time (i.e., same composition, exposure, etc) can be combined together to form a much sharper image. This technique is used in several areas, including astrophotography. Of course if you only have 1 picture to work with, well that's all the data you have to operate with and while you may be able to sharpen the image, there is always a trade-off, and you will loose quality in another aspect of the image.

    nntp://free.teranews.com/rec.photo.digital/<SoVtd.23$>

    <> wrote in message
    news:...
    > Isn't it possible to ........treat the camera as a black box?
    > Jos
    >


    No! You only have the information from the distorted picture.
    --
    For Welsh Military Flying visit .......
    www.groups.yahoo.com/group/V-A-S/



    [rec.photo.digital]
     
    Bill The Cat, Dec 9, 2004
    #10
  11. Mark² Guest

    "Bill The Cat" <DONT_SPAM_ME@tomk_AT_inforamp.net> wrote in message
    news:1102598602.373ce7a1117565c705c6eb9efb473c2a@teranews...
    > This is not true, or at least there is an exception. Several blurred

    images of the same subject taken from the same position at the same time
    (i.e., same composition, exposure, etc) can be combined together to form a
    much sharper image. This technique is used in several areas, including
    astrophotography. Of course if you only have 1 picture to work with, well
    that's all the data you have to operate with and while you may be able to
    sharpen the image, there is always a trade-off, and you will loose quality
    in another aspect of the image.
    ------------------------------------------
    A related technique is also used by some film scanners to boost dynamic
    range in a single image--by scanning it multiple times and combining the
    information within one image file.

    On the other hand, neither of our examples have much to do with what was
    originally asked... :)

    >
    >

    nntp://free.teranews.com/rec.photo.digital/<SoVtd.23$>
    >
    > <> wrote in message
    > news:...
    > > Isn't it possible to ........treat the camera as a black box?
    > > Jos
    > >

    >
    > No! You only have the information from the distorted picture.
    > --
    > For Welsh Military Flying visit .......
    > www.groups.yahoo.com/group/V-A-S/
    >
    >
    >
    > [rec.photo.digital]
     
    Mark², Dec 9, 2004
    #11
  12. Martin Brown Guest

    Mark² wrote:
    > "P.R.Brady" <> wrote in message
    > news:...
    >
    >> wrote:
    >>
    >>>Isn't it possible to somehow measure the distortion for a particular
    >>>lense, using a known calibration pattern that is photographed and then
    >>>apply a system-theory approach, i.e. treat the camera as a black box?
    >>>Jos

    >>
    >>I understood that some techniques like this were used for early Hubble
    >>images, but they still had an expensive trip to fit a correcting lens!

    >
    > Which indicates that it doesn't work well...


    Not at all. They work quite well. And they are still used to refine and
    enhance some of the best images taken with the new optics.

    The methods are good enough for the defects in the original Hubble
    optics to be computed from the initial blurred images. BUT it typically
    takes 2-3 orders of magnitude more work to process each image. That is
    an immense computational effort for every image.

    It isn't quite as good as doing it right in the first place but with a
    unique or valuable image it may be worthwhile.

    Regards,
    Martin Brown
     
    Martin Brown, Dec 9, 2004
    #12
  13. In article <1102598602.373ce7a1117565c705c6eb9efb473c2a@teranews>,
    "Bill The Cat" <DONT_SPAM_ME@tomk_AT_inforamp.net> writes:

    >This is not true, or at least there is an exception. Several blurred
    >images
    >of the same subject taken from the same position at the same time
    >(i.e., same
    >composition, exposure, etc) can be combined together to form a much
    >sharper image. This technique is used in several areas, including
    >astrophotography.
    >Of course if you only have 1 picture to work with, well that's all the data
    >you have to operate with and while you may be able to sharpen the image,
    >there is always a trade-off, and you will loose quality in another aspect
    >of the image.


    Mathematicians might locate a point object in an image,
    characterize the point object's 'signature' in the
    blurred/unfocused image, and create an 'inverse'
    (i.e., Wiener) filter to deblur/focus the image.

    Wiener filters usually have the undesired side effect of
    adding relatively high frequency noise to an image. A
    compromise inverse filter can introduce less high frequency
    noise at the cost of imperfect image restoration.

    <snip>

    'Hope that helps.

    Richard Ballard MSEE CNA4 KD0AZ
    --
    Consultant specializing in computer networks, imaging & security
    Listed as rjballard in "Friends & Favorites" at www.amazon.com
    Last book review: "Guerrilla Television" by Michael Shamberg
     
    Richard Ballard, Dec 9, 2004
    #13
  14. Bruce Murphy Guest

    writes:

    > Isn't it possible to somehow measure the distortion for a particular
    > lense, using a known calibration pattern that is photographed and then
    > apply a system-theory approach, i.e. treat the camera as a black box?
    > Jos


    Theoretically, sure. Unfortunately, you need to know startlingly large
    amounts of information about the lens *and* have a perfect image
    recorded.

    Don't bother.

    B>
     
    Bruce Murphy, Dec 9, 2004
    #14
  15. Bruce Murphy Guest

    (Richard Ballard) writes:

    > Mathematicians might locate a point object in an image,
    > characterize the point object's 'signature' in the
    > blurred/unfocused image, and create an 'inverse'
    > (i.e., Wiener) filter to deblur/focus the image.


    Unfortunately, the point-spread-function isn't going to be uniform
    over the field (or even in depth!) so things get very exciting, to say
    nothing of the horror of noise and distortion in the recorded image.

    Don't listen to the astronomy people, they just don't comprehend
    concepts like near objects :)

    B>
     
    Bruce Murphy, Dec 9, 2004
    #15
  16. angus Guest

    wrote:
    > Isn't it possible to somehow measure the distortion for a particular
    > lense, using a known calibration pattern that is photographed and then
    > apply a system-theory approach, i.e. treat the camera as a black box?
    > Jos
    >


    You might try either the free Focus Magic demo or Image Anayzer to see
    how well deconvolution methods work.

    With luck and patience these programs occasionally work.

    Angus
     
    angus, Dec 9, 2004
    #16
  17. Guest

    Bruce Murphy <> writes:

    > writes:


    >> Isn't it possible to somehow measure the distortion for a
    >> particular lense, using a known calibration pattern that is
    >> photographed and then apply a system-theory approach, i.e. treat
    >> the camera as a black box?


    > Theoretically, sure. Unfortunately, you need to know startlingly
    > large amounts of information about the lens *and* have a perfect
    > image recorded.


    It has been done, to recover some of the shots taken on one of the
    Apollo missions. The first photo of all of the earth was taken with
    the lens focused on 3 feet :( As that lens and camera returned, it was
    possible to construct a holographic filer to recover the image. An
    out of focus images has all of the in focus image info plus the out of
    focus image(s).

    One of the engineers comented that it was marginally cheaper tham
    re-flying the mission...

    BTW, the 32 aniversery of the last man on the moon was a few days ago.

    --
    Paul Repacholi 1 Crescent Rd.,
    +61 (08) 9257-1001 Kalamunda.
    West Australia 6076
    comp.os.vms,- The Older, Grumpier Slashdot
    Raw, Cooked or Well-done, it's all half baked.
    EPIC, The Architecture of the future, always has been, always will be.
     
    , Dec 9, 2004
    #17
  18. <> wrote in message
    news:...
    >
    > Isn't it possible to somehow measure the distortion for a particular
    > lense, using a known calibration pattern that is photographed and then
    > apply a system-theory approach, i.e. treat the camera as a black box?
    > Jos


    This is a big topic in astronomy. The problem is that even if you know the
    point-spread function of the lens, the solution is very unstable (i.e.,
    extremely tiny errors from film grain or the like will throw off the
    solution considerably). Look up "maximum-entropy deconvolution" for more
    about this.

    Astronomers have one advantage. An image of a star is a very good
    indication of the point-spread function. In ordinary photography, we don't
    have star-tests in every picture the way the astronomers do!


    --
    Clear skies,

    Michael A. Covington
    Author, Astrophotography for the Amateur
    www.covingtoninnovations.com/astromenu.html
     
    Michael A. Covington, Dec 9, 2004
    #18
  19. Annika1980 Guest

    >From:

    >It has been done, to recover some of the shots taken on one of the
    >Apollo missions. The first photo of all of the earth was taken with
    >the lens focused on 3 feet :( As that lens and camera returned, it was
    >possible to construct a holographic filer to recover the image. An
    >out of focus images has all of the in focus image info plus the out of
    >focus image(s).


    Now that would be an interesting feature to include in the next Canon. The RAW
    data would include data about distance to subject and the firmware would
    automatically know how to apply the focus correction.
     
    Annika1980, Dec 10, 2004
    #19
  20. Bruce Murphy Guest

    writes:

    > Bruce Murphy <> writes:
    >
    > > writes:

    >
    > >> Isn't it possible to somehow measure the distortion for a
    > >> particular lense, using a known calibration pattern that is
    > >> photographed and then apply a system-theory approach, i.e. treat
    > >> the camera as a black box?

    >
    > > Theoretically, sure. Unfortunately, you need to know startlingly
    > > large amounts of information about the lens *and* have a perfect
    > > image recorded.

    >
    > It has been done, to recover some of the shots taken on one of the
    > Apollo missions. The first photo of all of the earth was taken with
    > the lens focused on 3 feet :( As that lens and camera returned, it was
    > possible to construct a holographic filer to recover the image. An
    > out of focus images has all of the in focus image info plus the out of
    > focus image(s).


    It should be noted here that, as in the astrophotography area,
    everything in the image was within a poofteenth of infinity. Doing the
    same with a field full of things at different distances would be far
    harder.

    B>
     
    Bruce Murphy, Dec 10, 2004
    #20
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Valued User
    Replies:
    0
    Views:
    1,245
    Valued User
    Jun 28, 2003
  2. mushu
    Replies:
    3
    Views:
    603
    Guest
    Jan 11, 2004
  3. Bob Fusillo

    Can Smart Media cards be repaired?

    Bob Fusillo, Aug 16, 2004, in forum: Digital Photography
    Replies:
    16
    Views:
    5,084
    Jerry G.
    Aug 19, 2004
  4. Don Kirkman

    New, Hoping for Newsgroup Focused on Scanning and Editing

    Don Kirkman, Feb 16, 2008, in forum: Digital Photography
    Replies:
    14
    Views:
    512
    Peter Jason
    May 16, 2008
  5. Joe Mastroianni

    Nikon D5000 - how to make the darn thing STAY on center focused?

    Joe Mastroianni, Feb 22, 2013, in forum: Digital Photography
    Replies:
    33
    Views:
    610
    PeterN
    Mar 6, 2013
Loading...

Share This Page