If hue or saturation is what you are interested in consider using HSL or
HSV colour space. Converting the corectly registerd RGB images to this
is simple and available in most image editing softwares Photoshop, Gimp,
and ImageJ all have functions for this. Then finding the area of a
certain hue or saturation is a simple mater of thresholding the channel
of interest and measuring the resulting area.
/Ricardo Figueroa
Michelle Peckham wrote:
> Thanks Jerry
>
> These are pictures taken with a digital camera (i.e. Nikon or Canon) in
> 'raw' format. So your advice is great.
>
> This person wants to work out how much area is a particular shade/hue, so we
> thought grey scale numbers in R,G and B channels would be the thing to go
> for.
>
> So thanks for this.
>
> Michelle
>
>
> On 13/02/2009 16:46, "Jerry Sedgewick" <[log in to unmask]> wrote:
>
>
>> Is the file in the Raw format, such as those generated by Nikon and
>> Canon consumer cameras? Or did you mean that the person was asking about
>> the original (unaltered or raw) file?
>>
>> If it is the Nikon or Canon Raw format, then it was derived from a
>> mosaic chip and colors will have to be assigned to each pixel through
>> some kind of conversion "engine." Photoshop uses its own Adobe Camera
>> Raw (ACR) engine in all the CS versions (as I recall). If the lighting
>> for the study was consistent, and if the study occurred over more than
>> one imaging session, then the settings in Adobe Camera Raw should be set
>> similarly for the experiment, or the colors will have to be adjusted by
>> using a reference standard (such as a neutral gray card, Macbeth color
>> chart, or some consistent part of the specimen that is white or gray).
>> If the colors are not assigned or adjusted to a standard, then the
>> colors will be inconsistent from session to session.
>>
>> Canon and Nikon both have proprietary engines in which colors are
>> generated at each x, y position. These can also be used, but make sure
>> the gamma is set to 1, and, again, adjust colors based upon reference
>> standards if images were taken at more than one session.
>>
>> Before opening the image in Photoshop, the Color Settings in the menu
>> (Preferences>Color Settings) are changed from the RGB default setting
>> (sRGB....). A "CIE 1931 D65" setting is chosen because this keeps the
>> gamma at 1 (no alteration of post-ACR values). This color setting can be
>> downloaded from quickphotoshop.com, but make sure to click on Book:
>> Supplemental Info (not Download). Photoshop changes the gamma when using
>> any of the other color settings. Gamma is a reference to the adjustment
>> of pixel values destined for display on a computer monitor in which the
>> Greek symbol gamma is the exponent in a math formula.
>>
>> Once in Photoshop, in the menu, under Windows, choose Channels
>> (Windows>Channels). In the Channels palette, click on the upper right
>> arrowhead and choose Split Channels from the drop down list. This will
>> separate the red, green and blue channels for subsequent measurement.
>>
>> In the distant past, I believe I used Image J to generate pixel values
>> at each x, y position, but maybe it was MatLab or MetaMorph. Maybe
>> another person on the list can add that information.
>>
>> Cheers!
>>
>>
>>
>> Michelle Peckham wrote:
>>
>>> A little bit off the topic, but
>>>
>>> Someone here was asking me about analysing a ŚRAWą file for colour
>>> information, and the simplest way to do what she wants would be to
>>> extract the grey scale values out for the red, green and blue
>>> channels, and analyse them in a spread sheet.
>>>
>>> Does anyone know how to do this?
>>>
>>> Michelle
>>>
>
>
|
