| Re: how to make a mask from a picture and how to put... [message #38409 is a reply to message #38404] |
Thu, 04 March 2004 09:25   |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Thomas Nehls writes:
> thank you that is interesting, but:
>
> I especially want to avoid drawing by hand! I already have my black and
> white image bw (1,400,400) which results by some calclations from the
> original image "org"(3,400,400).
> Now I want to multiply these two images in that way, that everywhere
> where my "bw" is black (or white, whatever) should be black in the
> resulting image "res". the following loop did not work...
> for x=400, Y=400
> for X = 1:400
> for Y =1:400
> res(Z=0:3,X,Y) = bw(X,Y) * org(Z=0:3,X,Y)
> end
> end
> end
> Do you understand what I want to do?
Yes, but I was hoping you would be able to read between
the lines a bit. (I don't know why I thought this. I
haven't had a bit of luck this week! But there you go,
an eternal optimist!)
OK, I would REFORM your B&W image into a 2D array, not
a 3D. You will just get confused with that extra
1 dimension hanging around. (Or, at least, I do.)
image = Reform(bw)
I'm going to assume black image pixels are 0, everything
else is something other than 0.
mask = image GT 0
Mask now contains a 1 where you want the "light" to shine
through and 0 where you want to block it. If your situation
is the other way around, subtract 1 from mask.
Now you have to apply the mask to the three image
planes. Let's rearrange your pixel interleaved image into
a band interleaved so we don't have that pesky 1 dimension
to deal with:
maskedImage = Transpose(res, [1,0,2])
Now apply the mask to each color plane:
FOR j=0,2 DO maskedImage[0,0,j] = maskedImage[0,0,j]*mask
If you *have* to have a pixel-interleaved image:
maskedImage = Transpose(maskedImage, [2,0,1])
TV, maskedImage, True=1
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting
Coyote's Guide to IDL Programming: http://www.dfanning.com/
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
