| Re: fast search [message #50852 is a reply to message #50819] |
Fri, 20 October 2006 17:37   |
m.goullant@gmail.com
Messages: 10
Registered: July 2006
|
Junior Member |
|
|
greg michael wrote:
> Hi Marie,
Hello once again Greg! :-)
> Here's an annotated version.
>
> I understand you're looking for the maximum z within the vicinity of
> every point, with a gradually increasing radius of the vicinity. I
> can't see what you're doing with this value, though - does it feed back
> into the set of points somehow? What's the result you're trying to get?
Z is a elevation value (sea level)
I need the calculate de altimetric difference (difZ) after for example
a morphological operation of erosion to all data:
difZ = z-newZ
with this result I can make some comparisons and point exclusions...The
main problem I have is when I apply this filtering and iterative
process to a great volume of data.
Imagine if I want to calculate a morphological Opening operation
(erosion following dilation)?
> Where do these points come from?
These points come from an ASCII file (.dat;.txt;.xyz), here a little
sample:
"X","Y","Z"
645107.178,4808512.652,382.900
645106.228,4808512.642,381.940
645104.798,4808512.492,378.220
645103.678,4808512.502,377.500
645100.329,4808511.973,366.170
645089.639,4808512.643,367.570
645097.298,4808512.602,375.630
(...)
> It would be simple to reduce my search code to 2-D - just remove the
> z-lines, change the distance calculation, and the binning line to
> b=bx+by*n_split. But I'm not sure if this is right way - it depends
> what your z-values mean.
>
> I've just realised that my later versions don't handle the case where
> the pair lie across a subdivision boundary - only the slower first
> version does that. Something to fix...
>
I will try to see in what way I can adapt the cut method to this
situation: where when I center the kernel in a point(i) I only
calculate euclidean distances to that point in the subvolume or
subvolumes that are inside of the kernel
Any idea will be welcome :-)
Thanks!
> many greetings,
> Greg
Thanks! Same
Marie
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
