New User - ImageJ use with Sea Surface Temperature Images

[D Kelly O'Day]

I'm a new ImageJ user interested in analysis of sea surface temperature (sst) images.

I can't seem to find much information on use of ImageJ for climate image analysis. Can anyone point me in the right direction for plug-ins, websites, books to jump start my learning curve?

I am particularly interested in comparing the differences between 2 images taken one month apart. I'd like to be able to show the change from month to month.

Here's a link to a typical sst anomaly image.

How do I go about "subtracting" one image's ssta from another?

Thanks for any help you can give me.

[David Webster]

Assuming that they are the same size, just use Process/Image Calculator and
select the Subtract option.

David Webster

On Mon, Feb 8, 2010 at 7:43 AM, D Kelly O'Day <[hidden email]>wrote:

> I'm a new ImageJ user interested in analysis of sea surface temperature
> (sst)
> images.
>
> I can't seem to find much information on use of ImageJ for climate image
> analysis. Can anyone point me in the right direction for plug-ins,
> websites,
> books to jump start my learning curve?
>
> I am particularly interested in comparing the differences between 2 images
> taken one month apart. I'd like to be able to show the change from month to
> month.
>
> Here's a
> http://www.osdpd.noaa.gov/data/sst/anomaly/2010/anomnight.2.8.2010.giflink
> to a typical sst anomaly image.
>
> How do I go about "subtracting" one image's ssta from another?
>
> Thanks for any help you can give me.
>
>
> --
> View this message in context:
> http://n2.nabble.com/New-User-ImageJ-use-with-Sea-Surface-Temperature-Images-tp4534913p4534913.html
> Sent from the ImageJ mailing list archive at Nabble.com.
>

[Gabriel Landini]

On Monday 08 February 2010 19:48:37 you wrote:
> On Mon, Feb 8, 2010 at 7:43 AM, D Kelly O'Day wrote:
> > I'm a new ImageJ user interested in analysis of sea surface temperature
> > (sst) images.
[...]
> > How do I go about "subtracting" one image's ssta from another?

> Assuming that they are the same size, just use Process/Image Calculator and
> select the Subtract option.

Hmm... no,  because one cannot apply arithmetic operations between false
colour gif images (which are palette mapped) and expect a sensible result.

You probably need to convert/recalibrate the false colour to greyscale based
on the scale provided and only then do operations between the new images.

Cheers,
G.
 

[Kischell, Eric R.]

 D Kelly O'Day,

>I am particularly interested in comparing the differences between 2
images
>taken one month apart. I'd like to be able to show the change from
month to month.

Make sure that both of your time-point images are registered.
Each specific pixel at time point #1 represents the same physical
location as the same pixel at time point #2.
If not, turboreg may be of use http://bigwww.epfl.ch/thevenaz/turboreg/.

I assume that both of your images use an 8-bit pseudo-color LUT (i.e.
not full-color RGB)?
If so, then David Webster's procedure should be OK.
Note, you will need to 'invent' a new colour mapping to help interpret
your difference images (difference of indices).
More exotic change detection methods may be found in the relevant GIS
literature.

Regards,
e.-
===  

-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of D
Kelly O'Day
Sent: Monday, February 08, 2010 9:44 AM
To: [hidden email]
Subject: New User - ImageJ use with Sea Surface Temperature Images

I'm a new ImageJ user interested in analysis of sea surface temperature
(sst)
images.

I can't seem to find much information on use of ImageJ for climate image
analysis. Can anyone point me in the right direction for plug-ins,
websites,
books to jump start my learning curve?

I am particularly interested in comparing the differences between 2
images
taken one month apart. I'd like to be able to show the change from month
to
month.

Here's a
http://www.osdpd.noaa.gov/data/sst/anomaly/2010/anomnight.2.8.2010.gif
link
to a typical sst anomaly image.

How do I go about "subtracting" one image's ssta from another?

Thanks for any help you can give me.


--
View this message in context:
http://n2.nabble.com/New-User-ImageJ-use-with-Sea-Surface-Temperature-Im
ages-tp4534913p4534913.html
Sent from the ImageJ mailing list archive at Nabble.com.

[David Webster]

Hmm! Didn't think about the psuedo color as I assumed a scalar temperature
was the real variable. So, a conversion is needed if the scalar temp images
aren't available.

David Webster

On Mon, Feb 8, 2010 at 1:59 PM, Gabriel Landini <[hidden email]>wrote:

> On Monday 08 February 2010 19:48:37 you wrote:
> > On Mon, Feb 8, 2010 at 7:43 AM, D Kelly O'Day wrote:
> > > I'm a new ImageJ user interested in analysis of sea surface temperature
> > > (sst) images.
> [...]
> > > How do I go about "subtracting" one image's ssta from another?
>
> > Assuming that they are the same size, just use Process/Image Calculator
> and
> > select the Subtract option.
>
> Hmm... no,  because one cannot apply arithmetic operations between false
> colour gif images (which are palette mapped) and expect a sensible result.
>
> You probably need to convert/recalibrate the false colour to greyscale
> based
> on the scale provided and only then do operations between the new images.
>
> Cheers,
> G.
>
>

[Teresa W]

http://www.cooa.unh.edu/tutorial.jsp

has a great ImageJ and sea temp beginners ImageJ tutorial

Teresa

[Rodrigo Gonçalves-4]

Dear Kelly, Gabriel and others,

I have a similar situation but with other fluid: the atmosphere; and images representing ozone concentrations, for example from this image:
http://www.nasa.gov/images/content/160657main_OZONE_large.jpg

which uses the same palette as this image:
http://saferenvironment.files.wordpress.com/2008/12/ozone_hole.jpg

But, as Gabriel has pointed out, since this is a palette maped image, I guess it would be necessary to "convert/recalibrate the false colour to grayscale". Otherwise when I make a Plot Profile I get a plot which looks like a ascending zigzag line (plot attached).

My question is, how can this conversion/recalibration be performed without loosing information?

Thanks a lot to everybody that shares their knowledge and time in this list.

Rodrigo

________________________
Dr. Rodrigo J. Goncalves
Estación de Fotobiología Playa Unión
www.efpu.org.ar
________________________

--- El lun, 2/8/10, Gabriel Landini <[hidden email]> escribió:

De: Gabriel Landini <[hidden email]>
Asunto: Re: New User - ImageJ use with Sea Surface Temperature Images
A: [hidden email]
Fecha: lunes, 8 de febrero de 2010, 09:59 pm

On Monday 08 February 2010 19:48:37 you wrote:
> On Mon, Feb 8, 2010 at 7:43 AM, D Kelly O'Day wrote:
> > I'm a new ImageJ user interested in analysis of sea surface temperature
> > (sst) images.
[...]
> > How do I go about "subtracting" one image's ssta from another?

> Assuming that they are the same size, just use Process/Image Calculator and
> select the Subtract option.

Hmm... no,  because one cannot apply arithmetic operations between false
colour gif images (which are palette mapped) and expect a sensible result.

You probably need to convert/recalibrate the false colour to greyscale based
on the scale provided and only then do operations between the new images.

Cheers,
G.
 



      ____________________________________________________________________________________
¡Obtén la mejor experiencia en la web!
Descarga gratis el nuevo Internet Explorer 8.
http://downloads.yahoo.com/ieak8/?l=e1

[Gabriel Landini]

On Thursday 11 Feb 2010  15:25:11 you wrote:
> I have a similar situation but with other fluid: the atmosphere; and images
>  representing ozone concentrations, for example from this image:
>  http://www.nasa.gov/images/content/160657main_OZONE_large.jpg
>
> My question is, how can this conversion/recalibration be performed without
>  loosing information?
>
> Thanks a lot to everybody that shares their knowledge and time in this
>  list.

I would try to get the original data rather than reverse engineer from graphs.
One does not really know what type of processing the images have, and so the
quality of the data is unknown (as will be any conclusions one draws from
them).
The images above have - on top of the lut - some shading on one side of the
sphere.
And I would never use jpegs for this kind of job, since one has no control
over the compression artifacts.

Cheers,
G.