Posted by
Jerry (Gerald) Sedgewick on
Feb 18, 2016; 12:58pm
URL: http://imagej.273.s1.nabble.com/Color-Correction-for-Underwater-Photos-tp5015660p5015674.html
For getting accurate color I would follow methods used in the paint,
textile, photography, graphic arts, and a host of other industries. These
have been developed by color scientists with greater degrees of accuracy.
Rather than stay in RGB color space, colors are re-mapped in three
dimensional XYZ space.
This requires that you use a color standard such as the x-rite passport
chart and then calibrate colors to that standard. It's best if you work with
a RAW image, such as the those created by consumer cameras like those made
by Canon, Nikon, etc. In that instance you have options for software, but
the typical method is to use the Camera Raw program bundled with Photoshop.
You create a profile in the x-rite software, and then apply the profile in
the Camera Raw program. From there you can analyze image in Image J if you
wish. I suspect you're looking at changes in color. You may want to consider
measuring as they do in the color science world by looking at Delta E for
color variation.
This will get you 90 percent of the way there. Coral has specific reflection
properties that would require a color chart specifically made for Coral in
water. You can go one step further to develop your own profile by inviting
coral experts to weigh in on color just as was once done to develop the
means to assign numbers to color (LAB colors). I've done this for color of a
type of metal.
Not that I have a commercial affiliation with Adobe or X-rite, but your
application requires the right tool. Photoshop CC can be "rented" for about
10 dollars a month. In that program you open the accompanying program
Bridge, and then in Bridge you open images in Camera Raw to apply the
profile from the color chart. I've done several tests and have found this
method to be very consistent and repeatable.
Jerry Sedgewick
-----Original Message-----
From: John Clark
Sent: Wednesday, February 17, 2016 4:08 PM
To:
[hidden email]
Subject: Re: Color Correction for Underwater Photos
Only on rare occasions tune into ImageJ issues, as I’m not working in image
processing any more…
But… there is in ordinary camera work, either stills or motion pictures,
where one shoots some color chart with a certain lighting condition, in
order to generate a LUT for ‘correcting’ due to color casts of light.
Photoshop is abysmal for this. For some reason ‘still’ photographers don’t
seem to typically do this… they tend to select a ‘lighting’ option on their
camera Sun, Tungsten, Florescent, Cloudy… etc. and leave it at that.
If there is any residual cast, they ‘correct’ on the fly using photshop or
these days Lightroom widgetry.
In the motion picture world there is a better understanding of this, but
even so, creating a custom LUT is sort of a pain. The only tool I’ve found
where it is ‘dead simple’, is the Davinci Resolve package from Blackmagic
Designs.
The ‘pro’ package is expensive… but fortunately they have a ‘free’ version
which has this correction capability.
You could load ‘stills’ in as video clips, use the LUT tool to analyze your
chart and it would then develop the custom values required… provided you
used one of several industry standard charts.
I use a chart called “Spyder Color Checkr”, but there is support for another
chart called X-rite, from X-rite Photo.
Here is a table of ‘chart colors’ to ‘R,G,B’ values:
http://xritephoto.com/ph_product_overview.aspx?ID=824&Action=Support&SupportID=5159One would have to manually place a window over a color swatch, adjust gain,
hue, saturation, to achieve that color… like wise for several others… Having
never done this by hand, I don’t know how tedious or ‘circular’ this
would be, but at some point one should be able to ‘dial out’ the bluish cast
of the scattered light at some depth.
John Clark.
On Feb 17, 2016, at 8:19 AM, klesneski <
[hidden email]> wrote:
> Hi everyone,
> I am doing some analyses of tissue color of corals that I took with an
> underwater camera. I am trying to see if there are differences in tissue
> color among colonies. My idea is to use the color histograms of the
> selected
> colonies. In each image, I have a color standard card, placed next to the
> coral colony. One of the reasons why we include this card is because red
> light is attenuated with depth, so objects and images appear more blue the
> deeper you are, and the coral colonies are all at different depths. We
> know
> the RBG values of the color on the card. I am wondering if anyone has
> experience or advice on how I could use this color standard card in ImageJ
> to perform a sort of correction before using color histograms?
> Many thanks,
> KL
>
>
>
>
> --
> View this message in context:
>
http://imagej.1557.x6.nabble.com/Color-Correction-for-Underwater-Photos-tp5015660.html> Sent from the ImageJ mailing list archive at Nabble.com.
>
> --
> ImageJ mailing list:
http://imagej.nih.gov/ij/list.html--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html
--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html