| Re: Point within country boundary [message #87334 is a reply to message #87330] |
Sat, 25 January 2014 04:54   |
Fabzi
Messages: 305
Registered: July 2010
|
Senior Member |
|
|
Hi,
On 24.01.2014 22:36, David Fanning wrote:
> Well, believe it or not, this is nearly instantaneous!
Contain points is musch slowier than computeMask, see following test:
pro test_contain_points
poly_x = [10, 90, 90, 10, 10] + 0.1
poly_y = [10, 10, 90, 90, 10] + 0.1
n = 1800
o = IDLanROI(poly_x, poly_y)
print, 'Compute mask'
tic
result = o->ComputeMask(DIMENSION=[n,n])
toc
i = Image(result)
xx = INDGEN(n) # (LONARR(n) + 1)
yy = INDGEN(n) ## (LONARR(n) + 1)
mask = BYTARR(n,n)
print, 'Contain point'
tic
result = o->ContainsPoints(xx,yy)
toc
mask[where(result)] = 255
i = Image(mask)
mask = BYTARR(n,n)
print, 'This does the trick'
tic
totest = where(o->ComputeMask(DIMENSION=[n,n]))
result = o->ContainsPoints(xx[totest],yy[totest])
toc
mask[totest[where(result)]] = 255
i = Image(mask)
end
on my machine:
IDL> test_contain_points
Compute mask
% Time elapsed: 0.00040602684 seconds.
Contain point
% Time elapsed: 0.53887701 seconds.
This does the trick
% Time elapsed: 0.0035710335 seconds.
Since contains points can do what compute mask can't, I use the trick
above to spare computing time...
Cheers,
Fabien
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
