| simple vectorizing problem [message #62465] |
Mon, 08 September 2008 18:47 |
dpm314
Messages: 3
Registered: September 2008
|
Junior Member |
|
|
Hey - here is a vectorizing question from an experienced programmer
'newish' to IDL.
I have a stack of images dimensions (nx,ny, num_images) , a LOT of
data sometimes, on which I need to preform a whole mathematical
operation on. For example, one thing I need to do is create an array
of the mean of each image.
I've done this successfully with something like
averages = fltarr(num_images)
for k = 0, NUM_OF_CLICKS-1 do averages[k] = mean(image(*,*,k)
which makes perfect sense to a c/c++/Fortran programmer. But, this
really hangs up for the size images I am working with (512*512*70000
or more). Is there a way to vectorize? I've tried several things
like:
averages = fltarr(num_images)
a = indgen(num_images)
averages(a) = mean(image(*,*,a))
I've also tried reforming the image to by 2D, with dimensions [nx*ny,
num_images] and doing a similar thing, but no luck. That included
making an array of integers, call it b, where b = 0, nx*ny*1 - 1,
nx*ny*2 - 1, nx*ny*3 - 1 ... and doing something like
averages(a) = mean(image(b[a]:b[a+1], a)
(I know this probably will run out of bounds, I can't remember right
now what I did, but it still didn't work and I had the series worked
out right...)
It seems that in a language like IDL there should be a way to do
something like this without writing a for-loop.
Thanks,
David M.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
