| Re: Polygon Clipping Algo in IDL [message #46722 is a reply to message #46699] |
Tue, 13 December 2005 15:02   |
Mark Hadfield
Messages: 783
Registered: May 1995
|
Senior Member |
|
|
JD Smith wrote:
> On Tue, 13 Dec 2005 14:38:42 +1300, Mark Hadfield wrote:
>
> Hey Mark... remind me what your version did different from mine? I recall
> spending weeks trying to vectorize this algorithm, only to give up and
> write it as a DLM. It's fairly slow in IDL. Lately I've been exploring
> the clipping library GPC, linking it as a DLM.
It's been a while. Didn't your POLYCLIP clip to a square? Whereas my
routines clip to a single line (must be parallel to the X or Y axis for
MGH_POLYCLIP and of arbitrary orientation for MGH_POLYCLIP2). Clipping
to a square or rectangle therefore requires 4 applications. I settled on
the line-clipping functions because they are more general and applying
them 4 times is not significantly slower that applying a
rectangle-clipping form once.
I spent a fair bit of time tweaking the code and achieved a modest
speed-up over the code you sent me (perhaps 2x). The thing that made the
difference was converting a couple of functions into in-line code.
I actually wanted this code for the following situation: I have a 2D
grid of cells (usually rectangular, sometimes curvilinear) and some
polygons which, for the sake of argument, we can consider to be land
areas. For each cell in the grid, i want the fraction that is inside one
of the polygons. So for each cell I clip every polygon and calculate the
area (if any) in the clipped polygon. This approach is *not*
particularly fast and implementing the polygon clipping in IDL slows it
down further. I'm sure DLM-ed C code would be faster and am interested
in your results.
--
Mark Hadfield "Kei puwaha te tai nei, Hoea tahi tatou"
m.hadfield@niwa.co.nz
National Institute for Water and Atmospheric Research (NIWA)
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
