Are there any programs that can convert color infra-red photos to actual color?

Discussion in 'Digital Photography' started by Scotius, Jul 16, 2010.

  1. Scotius

    Scotius Guest

    I know that color infra-red images look really weird (for lack
    of a better term), but I once read that infra-red light cuts through
    fog/haze etc better than regular light, which I suppose is why B & W
    infra-red shots always look better than B & W shots without IR flash.
    So I'm wondering if there's a program that could accurately
    predict based on IR color what the colors present should be, and
    convert them, so it would be possible to do color shots better in
    haze, etc.
    Anyone know of anything like this?
     
    Scotius, Jul 16, 2010
    #1
    1. Advertisements

  2. Scotius

    Guest Guest


    I can't see how it might work. The color information you are
    looking for is not in the data you have.

    I would guess it may be possible to make the results look a
    little less odd, but I doubt that it would help any.

    It would be something like trying to make a Big Mac taste like
    cheese cake with cherry topping with out having cheese or cherries or
    even knowing that what you have to start with is a Big Mac.
     
    Guest, Jul 16, 2010
    #2
    1. Advertisements

  3. Could you tell what colour a particular car in a B&W photo is? That's
    essentially the same question. In a B&W image you have lost the colour
    information. In an infra-red image you have also lost colour
    information. Nothing will recover what you didn't record. You can
    guess, and paint it any colour you like, just as they used to do with
    B&W photos.
    Only similar questions. I was once asked by a senior member of the UK
    Royal Family why the false colour pictures from a thermal camera,
    representing temperature from blue being cold to red being hot, made
    someone's shirt look orange when it was obviously blue. Just as I
    repeated that it was false colour, a colleague jumped in and told him
    not to worry because I would have that fixed in a day or two. No such
    thing as a stupid question, just stupid people.
     
    Kennedy McEwen, Jul 16, 2010
    #3
  4. Scotius

    DanP Guest

    The information from an IR image has nothing to do with the colour.
    A hot green mug will look different in IR than a plant with the same
    shade of green.
    So you cannot map IR to visible colour.

    DanP
     
    DanP, Jul 16, 2010
    #4
  5. Scotius

    Nervous Nick Guest

    Why would you want to do this, even if it were at all possible?
     
    Nervous Nick, Jul 17, 2010
    #5
  6. []
    Be careful not to confuse near-IR with far-IR. With digital cameras and
    film it's the region just beyond the red end of the visible spectrum which
    people call "IR" - a wavelength of ~0.8um. Here, the prime difference is
    that the reflectance of vegetation is much higher and hence the
    characteristic appearance of monochrome IR images.

    Put briefly: to see thermal radiation from a hot mug requires an imager
    sensitive in the 10um region of the spectrum, which might require a cooled
    detector. To see fires of a few hundred degrees C, imagers sensitive to
    the 3-5um region of the spectrum work the best.

    Cheers,
    David
     
    David J Taylor, Jul 17, 2010
    #6
  7. This particular Royal was very savvy about the technology and commented
    that he had used something similar during his combat experience, to
    which another colleague remarked "I know". When asked by the Royal in
    question how he knew this, my colleague replied "Your mom told me!". ;-)
     
    Kennedy McEwen, Jul 17, 2010
    #7
  8. No it doesn't, 3-5um works just as well - better in fact.
    Not at 10um it doesn't.

    The radiation from the mug may peak around 10um, but with modern cooled
    sensors its the radiation contrast that determines the optimum band. The
    contrast at mug-like temperatures in the 3-5um region is about 4%/degC
    and roughly double what it is at 10um.
     
    Kennedy McEwen, Jul 17, 2010
    #8
  9. I was trying to keep it simple! Note that I said "might require a cooled
    detector", not "will". Accepted that the contrast is greater, but you
    then quote the use of a cooled detector.

    All space-based instruments which I know about use cooled detectors, for
    looking at earth-like (i.e. mug-like) temperatures, and the NET is
    typically lower in a 10um detector than in a 3-5um one.

    The point remains to distinguish between far-IR and near-IR uses of the
    term "infra-red" (and those tend not to be hard and fast terms either).

    Cheers,
    David
     
    David J Taylor, Jul 17, 2010
    #9
  10. Trying being the operative word.
    But you said "requires...10um", it doesn't and if you choose to work at
    10um it certainly doesn't require a cooled detector.
    I assume that is because they are "looking" at each spot for a limited
    time, rather than operating to full performance without time
    limitations. Modern cooled infrared detectors are more like the sensors
    in digital cameras, matrix arrays. In these devices the exposure time
    is limited by the well capacity (just as the exposure of digital camera
    sensors to avoid saturation is limited by the pixel storage capacity).
    When you are storage capacity limited, instead of time limited, as with
    modern high performance cooled sensors, the NET of 3-5um detectors is
    around half that of an equivalent 10um device. For example, two
    otherwise identical devices, well into BLIP performance:
    http://tiny.cc/wnyeb is a 3-5um device with NETD of ~17mK while
    http://tiny.cc/f0032 is its 10um counterpart with NETD of ~32mK.

    Or, in exactly the same device which is sensitive in spectral regions:
    http://tiny.cc/ni4dw when operating dedicated to each band the 3-5um
    NETD is ~11mK, whilst the 8-10um NETD is ~22mK.
     
    Kennedy McEwen, Jul 17, 2010
    #10
  11. To avoid exactly this level of detail.
    Yes, I'm talking scanned rather than staring arrays.

    Thanks for the further info.

    Cheers,
    David
     
    David J Taylor, Jul 17, 2010
    #11
  12. Scotius

    Peter Guest


    Almost!
    Chlorophyll is transparent to waves in the near IR spectrum. Hence it
    appears to be white. Similarly many clothing dyes are transparent in that
    spectrum and also appear white. Particles in the air cause the sky to appear
    blue. In this spectrum these particles do not reflect the energy waves and
    thus appear black.

    BTW I have submitted an IR photo in the SI.
     
    Peter, Jul 18, 2010
    #12
  13. []
    ... and with resolutions today of 3712 x 3712 pixels, and in a few years
    time of nearer 11,000 pixels square. Mechanically scanned, with mirror
    optics, covering 0.6um to 13um with a set of detectors.

    Cheers,
    David
     
    David J Taylor, Jul 18, 2010
    #13
  14. []
    []

    If anyone would like to know a little more, there's a training module
    here, entitled:
    "Monitoring Vegetation From Space".

    http://www.satreponline.org/landsaf/index.htm

    This section may be of particular interest:

    http://www.satreponline.org/landsaf/navmenu.php?page=3.0.0

    Cheers,
    David
     
    David J Taylor, Jul 18, 2010
    #14
  15. Scotius

    Scotius Guest

    I never thought of it that way. Well, thanks for making me
    think... too much actually. :)
    The reason I was asking is that when I recently covered a
    concert for a local magazine, I spoke to one of the stage managers,
    and he told me it would be okay to take a few shots with flash, but
    not too many.
    I thought that maybe with infra-red (invisible to the naked
    eye), it would be great to be able to take a fully illuminated shot,
    then go home and process it on the computer.
    I suppose it would work fine it you were shooting B & W
    infra-red, just not color, which is what I was most interested in.
    Well, you've convinced me. I still think B & W infra-red is a
    good idea for not bothering a band or crew though... or am I wrong
    about that too? Oh God don't let me be wrong about that too... can an
    IR flash be seen (I'm hoping not)?
     
    Scotius, Jul 28, 2010
    #15
  16. Scotius

    Scotius Guest

    Alexander Shulgin would probably be interested in it if he was
    still alive... and then I'd be immediately NOT interested in it. I
    just can't respect acid-loving hippies.
     
    Scotius, Jul 28, 2010
    #16
  17. Scotius

    Scotius Guest

    It's probably not so much that as it is them being distracted
    by worrying if anyone's noticed their club feet.
     
    Scotius, Jul 28, 2010
    #17
  18. Scotius

    Scotius Guest

    I'm glad everyone is enjoying this discussion so much,
    although I was pretty much done when I found it that "...it wouldn't
    work" :)
    Meanwhile, I've now been convinced that I might be best off to
    stay well away from infra-red completely.
    Am I to understand that not only would I not be able to
    convert the infra-red colors back to "regular" color since there's a
    lack of information about the regular color in the infra-red image,
    but also that there are things out there that absorb the infra-red and
    make using it for photography far more difficult?
    I might to to B & W infra-red for photos where I don't want to
    distract anyone with flash, but I won't go anywhere near color
    infra-red... unless I want to take spooky looking pix that is.
     
    Scotius, Jul 28, 2010
    #18
  19. Scotius

    Scotius Guest

    I was recently covering a concert for a local magazine, and
    asked a stage manager about taking pix with the flash. He said go
    ahead and take a few with flash, but not too many, so as not to be
    distracting.
    I had read about B & W infra-red photography in an old issue
    of Popular Mechanics, I think, that my Dad had lying around somewhere.
    Then I had read an article on color infra-red, and I thought "Oh, well
    then I'll just shoot pix like that in color infra-red and convert them
    on the computer back at home. People can't see infra-red, so there
    won't be a visible flash, and I'll convert the pix and have great
    shots that didn't bother anyone".
    It's since been explained to me that there's no method of
    converting the color infra-red pix, since the information about actual
    color is just as gone in those as it would be in black and white.
    I suppose for a huge event I could take one with flash and
    then recolor manually and submit the pix a couple years later :), but
    that's not really what I was looking to be able to do.
     
    Scotius, Jul 28, 2010
    #19
  20. Scotius

    otter Guest

    It is often not necessary, or even desirable, to take pictures with
    flash at a concert. These were taken (not by me) without flash. I
    wouldn't say they are great, but it shows that it is at least
    possible:

    http://www.flickr.com/photos/nataliekbeats/4546074977/in/set-72157623793453107/
    http://www.flickr.com/photos/nataliekbeats/4546074927/in/set-72157623793453107/

    Those were taken with a Rebel Xsi, which I think is similar to your
    D3000, as far as sensor size.

    You could also get a fast prime lens, or perhaps even a camera with a
    FF sensor if you wanted better low-light performance.
     
    otter, Jul 29, 2010
    #20
    1. Advertisements

Ask a Question

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

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.