| Re: Reading an arbitrary profile from 2D FITS data [message #76716] |
Wed, 29 June 2011 04:07  |
Bringfried Stecklum
Messages: 75
Registered: January 1996
|
Member |
|
|
Balt wrote:
> Hi all,
>
> A seemingly very simple problem has me stumped: How do I extract into
> a vector a profile line from a 2D FITS data set? The IDL astro lib
> doesn't seem to contain such a function for 2D, only for 3D, and then
> only along a cardinal axis. That in turn is easy to do also but I need
> it to go from for example (x,y) 100,100 to 320,240.
>
> Any ideas?
>
> Cheers
>
> - Balt
Perhaps these few lines of code may help. Regards, Bringfried
; extract profile from 2D image
; input - image
; - profile x and y start/end coords [x0,x1,y0,y1]
; output - 1D profile
function profile,image,xy
; check index bounds here...
if (xy[0] eq xy[1]) then return,image[xy[0],xy[2]:xy[3]]
if (xy[2] eq xy[3]) then return,image[xy[0]:xy[1],xy[2]]
; non-trivial case
a=(xy[2]-xy[3])/(xy[0]-xy[1])
b=xy[2]-a*xy[0]
x=xy[0]+indgen(xy[1]-xy[0]+1)
y=round(a*x+b)
return,reform(image[[x],[y]])
end
|
|
|
|
| Re: Reading an arbitrary profile from 2D FITS data [message #76769 is a reply to message #76716] |
Mon, 04 July 2011 21:26  |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
On Jun 29, 7:07 am, Bringfried Stecklum <steck...@tls-tautenburg.de>
wrote:
> Balt wrote:
>> Hi all,
>
>> A seemingly very simple problem has me stumped: How do I extract into
>> a vector a profile line from a 2D FITS data set? The IDL astro lib
>> doesn't seem to contain such a function for 2D, only for 3D, and then
>> only along a cardinal axis. That in turn is easy to do also but I need
>> it to go from for example (x,y) 100,100 to 320,240.
>
>> Any ideas?
>
>> Cheers
>
>> - Balt
>
> Perhaps these few lines of code may help. Regards, Bringfried
>
> ; extract profile from 2D image
> ; input - image
> ; - profile x and y start/end coords [x0,x1,y0,y1]
> ; output - 1D profile
>
> function profile,image,xy
> ; check index bounds here...
> if (xy[0] eq xy[1]) then return,image[xy[0],xy[2]:xy[3]]
> if (xy[2] eq xy[3]) then return,image[xy[0]:xy[1],xy[2]]
> ; non-trivial case
> a=(xy[2]-xy[3])/(xy[0]-xy[1])
> b=xy[2]-a*xy[0]
> x=xy[0]+indgen(xy[1]-xy[0]+1)
> y=round(a*x+b)
> return,reform(image[[x],[y]])
> end
It's the right idea.
The original question leaves open to interpretation what a "vector
profile" might be. I would suspect it would be equi-spaced in 2D
space. In which case it might be best to make a unit vector pointing
from A to B, and then INTERPOLATE() along multiples of that unit
vector.
Something along these lines (completely untested)
; extract interpolated profile from 2D image (equispaced)
; input - image
; - profile x and y start/end coords [x0,x1,y0,y1]
; n - number of intervals between start/end point
; output - 1D profile (returns N+1 samples)
function profile,image,xy,n
xya = xy[0:1] ;; Start
xyb = xy[2:3] ;; End
; Distance from start to end
dist = sqrt(total( (xyb - xya)^2 ))
;; Degenerate cases
if n EQ 0 then return, image[xya[0], xya[1]]
;; Unit vector - there are N intervals from start to end
unit = (xyb - xya)/n
;; Uniformly sampled points in X and Y
xx = xya[0] + dindgen(n+1)*unit[0]
yy = xya[1] + dindgen(n+1)*unit[1]
;; Interpolate
return, interpolate(image, xx, yy)
end
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
