| Overlaying where data [message #58288] |
Thu, 24 January 2008 12:39  |
jtmcahill
Messages: 26
Registered: October 2007
|
Junior Member |
|
|
Hello,
So, I've got a tvscl image that I've done some analysis on. With the
where function I've found some places I want to highlight in that
image by overlaying those areas with some color. I've managed to do
something close with the tv function but it changes the color of the
original tvscl image and the highlighted area is a barely
distinguishable (washed out red color). Does anyone have any ideas on
how to do this. I'm not contouring or anything special, just found
significant area in an image and want to highlighted so I can point
and say "there is where there is a significant amount of X".
Thanks
|
|
|
|
| Re: Overlaying where data [message #58336 is a reply to message #58288] |
Mon, 28 January 2008 00:58  |
izimine
Messages: 11
Registered: November 1997
|
Junior Member |
|
|
> mask = (minerals gt val1 * minerals lt val2) + (minerals gt val3)
sorry, you will probably want to do
mask = (minerals gt val1) * (minerals lt val2) + (minerals gt val3)
|
|
|
|
| Re: Overlaying where data [message #58338 is a reply to message #58288] |
Sun, 27 January 2008 19:46 |
izimine
Messages: 11
Registered: November 1997
|
Junior Member |
|
|
> Ok, let's see if I can explain this more clearly. First, I display
> the original image in tvscl. Then, I've got a second array that I've
> determined the % of a given mineral per pixel (say from 0 to 1 or 0 to
> 100 either way you want to look at it). I can tvscl the % mineralogy
> no problem in a window on its own (colored or grey scale).
if you want to do it in direct graphics with color tables...
; load background and overlay colors (change ncolors as needed)
loadct, 0, ncolors=128
loadct, ???, bottom=128, ncolors=128 ; or use tvlct
; bytscale accordingly
back = bytscl(orig, top=127)
over = bytscl(minerals, top=127)+128
; make a binary mask of what you want to show
mask = (minerals gt val1 * minerals lt val2) + (minerals gt val3)
; display
tv, back*(1-mask) + over*mask
if you need transparency better look at object graphics
cheers
|
|
|
|
| Re: Overlaying where data [message #58360 is a reply to message #58288] |
Fri, 25 January 2008 12:53 |
pgrigis
Messages: 436
Registered: September 2007
|
Senior Member |
|
|
David Fanning wrote:
> jtmcahill@gmail.com writes:
>
>> Ok, let's see if I can explain this more clearly. First, I display
>> the original image in tvscl.
>
> If you are really using TVSCL to display your image, I think
> you are already in trouble or you will be in trouble soon.
> If colors matter to you, forget you ever heard anything about
> TVSCL. Learn how to use TV and BYTSCL, including *all* the BYTSCL
> keywords. You will be undermining a lot of the work we are doing
> here if you use TVSCL.
>
> (And if you *really* want to work in IDL, get TVIMAGE or IMGDISP
> from one of the usual places on the Internet. You don't want to
> be using TV either. :-)
>
>> Then, I've got a second array that I've
>> determined the % of a given mineral per pixel (say from 0 to 1 or 0 to
>> 100 either way you want to look at it). I can tvscl the % mineralogy
>> no problem in a window on its own (colored or grey scale). But what
>> I'd like to do is to overlay the original image that is tvscl, with
>> another tvscl (which is the % mineralogy) without effecting the
>> original image. So, it is similar to highlighting the area of the
>> image that fit my criteria (like above), but now I'd like it to
>> visually show the areas with a higher and lower % of that mineral as
>> well. The first image would be grey scale, the second overlayed image
>> probably in color. You may think that the entire image would be
>> colored, but no. Because I've already picked out pixels that fit
>> another geochemical criteria first. So, I only have ~20% of the
>> original image to cover. If I display the second image alone, the
>> observer has no context for what they are looking at. But, if I over
>> lay it on the first image, that will provide the context. That's what
>> I'm shooting for.
>
> Have a look at this article, I think this describes what you are
> after:
>
> http://www.dfanning.com/color_tips/color_overlay.html
Hi,
In this case, I'd argue for a different approach. Assuming that the OP
has
two images im1 (with color scale 1) and im2 (with color scale 2) and a
ROI, such that the goal is to have a plot of im1 (with colors 1)
outside
the ROI and im2 (with colors 2) inside the ROI, the he could proceed
in the following way:
1) convert im1 + color scale 1 to a true color image (call it im1true)
2) convert im2 + color scale 2 to a true color image (im2true)
3) substitute pixels of im1true inside roi with corresponding values
from im2true
4) device,/decomposed
4) tv,im1true,/true
That should not be too hard to implement... here's an example (not
polished
nor optimized, just a hint)
;example
;create 2 images
im1=dist(256,256)
im2=rebin(findgen(256),256,256)
;define ROI
ind=array_indices(im1,where(im1 GT 100))
;black-white is col scale for im1
loadct,0
tvlct,r,g,b,/get
im1truecol=im2truecol(im1,r,g,b);convert to true color
;red temp scale for im 2
loadct,3
tvlct,r,g,b,/get
im2truecol=im2truecol(im2,r,g,b);convert to true color
;assign roi pixel from 2 to 1
FOR i=0L,n_elements(ind)/2-1 DO BEGIN
im1truecol[ind[0,i],ind[1,i],*]=im2truecol[ind[0,i],ind[1,i] ,*]
ENDFOR
device,decomposed=1
tv,im1truecol,true=3
using the function (again, this is a bit rough and is not polished)
FUNCTION im2truecol,im,r,g,b
s=size(im)
sx=s[1]
sy=s[2]
n=n_elements(r)
maxim=max(im)
minim=min(im)
im=(im-min(im))/(max(im)-min(im))*n
im3col=[[[reform(r[reform(im,sx*sy)],sx,sy)]], $
[[reform(g[reform(im,sx*sy)],sx,sy)]], $
[[reform(b[reform(im,sx*sy)],sx,sy)]]]
return,im3col
end
Ciao,
Paolo
>
> Cheers,
>
> David
> --
> David Fanning, Ph.D.
> Fanning Software Consulting, Inc.
> Coyote's Guide to IDL Programming: http://www.dfanning.com/
> Sepore ma de ni thui. ("Perhaps thou speakest truth.")
|
|
|
|
| Re: Overlaying where data [message #58361 is a reply to message #58288] |
Fri, 25 January 2008 12:05 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
jtmcahill@gmail.com writes:
> Ok, let's see if I can explain this more clearly. First, I display
> the original image in tvscl.
If you are really using TVSCL to display your image, I think
you are already in trouble or you will be in trouble soon.
If colors matter to you, forget you ever heard anything about
TVSCL. Learn how to use TV and BYTSCL, including *all* the BYTSCL
keywords. You will be undermining a lot of the work we are doing
here if you use TVSCL.
(And if you *really* want to work in IDL, get TVIMAGE or IMGDISP
from one of the usual places on the Internet. You don't want to
be using TV either. :-)
> Then, I've got a second array that I've
> determined the % of a given mineral per pixel (say from 0 to 1 or 0 to
> 100 either way you want to look at it). I can tvscl the % mineralogy
> no problem in a window on its own (colored or grey scale). But what
> I'd like to do is to overlay the original image that is tvscl, with
> another tvscl (which is the % mineralogy) without effecting the
> original image. So, it is similar to highlighting the area of the
> image that fit my criteria (like above), but now I'd like it to
> visually show the areas with a higher and lower % of that mineral as
> well. The first image would be grey scale, the second overlayed image
> probably in color. You may think that the entire image would be
> colored, but no. Because I've already picked out pixels that fit
> another geochemical criteria first. So, I only have ~20% of the
> original image to cover. If I display the second image alone, the
> observer has no context for what they are looking at. But, if I over
> lay it on the first image, that will provide the context. That's what
> I'm shooting for.
Have a look at this article, I think this describes what you are
after:
http://www.dfanning.com/color_tips/color_overlay.html
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Sepore ma de ni thui. ("Perhaps thou speakest truth.")
|
|
|
|
| Re: Overlaying where data [message #58362 is a reply to message #58288] |
Fri, 25 January 2008 11:53 |
jtmcahill
Messages: 26
Registered: October 2007
|
Junior Member |
|
|
On Jan 25, 8:52 am, David Fanning <n...@dfanning.com> wrote:
> jtmcah...@gmail.com writes:
>> Great! Thanks! I had to download your library to get some of the
>> functions of tvlct to work but it works great. Now, I might be
>> pushing my luck here, but each data area has a range of values as
>> well. Basically, I modeled a multispectral image array and the areas
>> I'm highlighting have results of my modeling. If I want to, can I
>> overlay a false colored tvscl image in a similar manner?
>
> I don't understand the question, but I'm sure the answer
> is "yes", you can do whatever you like with IDL. :-)
>
> I don't know what "overlay a false colored tvscl image
> in a similar manner" means in the context of the discussion
> so far. Can you elaborate?
>
> Cheers,
>
> David
> --
> David Fanning, Ph.D.
> Fanning Software Consulting, Inc.
> Coyote's Guide to IDL Programming:http://www.dfanning.com/
> Sepore ma de ni thui. ("Perhaps thou speakest truth.")
Ok, let's see if I can explain this more clearly. First, I display
the original image in tvscl. Then, I've got a second array that I've
determined the % of a given mineral per pixel (say from 0 to 1 or 0 to
100 either way you want to look at it). I can tvscl the % mineralogy
no problem in a window on its own (colored or grey scale). But what
I'd like to do is to overlay the original image that is tvscl, with
another tvscl (which is the % mineralogy) without effecting the
original image. So, it is similar to highlighting the area of the
image that fit my criteria (like above), but now I'd like it to
visually show the areas with a higher and lower % of that mineral as
well. The first image would be grey scale, the second overlayed image
probably in color. You may think that the entire image would be
colored, but no. Because I've already picked out pixels that fit
another geochemical criteria first. So, I only have ~20% of the
original image to cover. If I display the second image alone, the
observer has no context for what they are looking at. But, if I over
lay it on the first image, that will provide the context. That's what
I'm shooting for.
Mahalo,
Josh
|
|
|
|
| Re: Overlaying where data [message #58366 is a reply to message #58288] |
Fri, 25 January 2008 10:52 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
jtmcahill@gmail.com writes:
> Great! Thanks! I had to download your library to get some of the
> functions of tvlct to work but it works great. Now, I might be
> pushing my luck here, but each data area has a range of values as
> well. Basically, I modeled a multispectral image array and the areas
> I'm highlighting have results of my modeling. If I want to, can I
> overlay a false colored tvscl image in a similar manner?
I don't understand the question, but I'm sure the answer
is "yes", you can do whatever you like with IDL. :-)
I don't know what "overlay a false colored tvscl image
in a similar manner" means in the context of the discussion
so far. Can you elaborate?
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Sepore ma de ni thui. ("Perhaps thou speakest truth.")
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
