[R-sig-Geo] random sampling along contour

[Edzer Pebesma]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Seth, after running your script, I tried:

b = gBuffer(t.sl, width = .1, byid=FALSE)
plot(b, col='#ff000088', add=TRUE)
i = gIntersection(b, v.sldf, byid = TRUE)
plot(i, col = 'blue', add=TRUE)
pts = spsample(i[1], n=1, type='random')
l = lapply(1:length(i), function(x) spsample(i[x], n=1, type='random'))
npts = do.call(rbind, l[!sapply(l, is.null)])
plot(npts, add=TRUE, col = 'green')

which goes a fixed distance from the transect, but not a fixed
distance along each contour. Note that you get one point per unique
contour level.


On 10/09/2013 03:01 PM, Seth Bigelow wrote:
> Barry,
> 
> I do mean exactly what you have stated, i.e., "plus or minus up to
> 50m either side of the points created by your plot, but along the
> contour line? In other words, I go to each of those points on your
> plot, choose a runif(1,-50,50) and traverse that distance on the
> contour line to get my sampling point?"
> 
> In my real-world application, there do not seem to be either of the
> 'edge' issues you raise, since the flank of the mountain I am
> sampling is not highly dissected. But if the contour line is a
> loop, it would be fine to 'go around', or just throw an 'NA'.
> Similarly, if the contour line ends, an 'NA' would be ok.
> 
> Thanks for thinking about this!
> 
> -Seth Bigelow
> 
> 
> -----Original Message----- From: b.rowlingson at gmail.com
> [mailto:b.rowlingson at gmail.com] On Behalf Of Barry Rowlingson Sent:
> Wednesday, October 09, 2013 8:18 AM To: Seth Bigelow Cc:
> r-sig-geo at r-project.org Subject: Re: [R-sig-Geo] random sampling
> along contour
> 
> On Wed, Oct 9, 2013 at 12:32 PM, Seth Bigelow <seth at swbigelow.net>
> wrote:
>> Hello all:
>> 
>> 
>> 
>> I wish to place sampling plots at regular elevation intervals up
>> a
> mountain.
>> Plot centers should be placed at random within a 100 m wide
>> transect along the contour. Can anyone suggest an efficient way
>> of doing this? Code below places plots on a single transect on
>> the volcano dataset, but does not do the 'along the contour'
>> randomization.
> 
> I'm not clear what you mean here, the plot is a big help though
> and +1 for that.
> 
> By "a 100m wide transect along the contour" do you mean plus or
> minus up to 50m either side of the points created by your plot, but
> along the contour line? In other words, I go to each of those
> points on your plot, choose a runif(1,-50,50) and traverse that
> distance on the contour line to get my sampling point?
> 
> What if the contour line at that point is a loop less than 100m
> long? Do I go around? The other edge case is literally an edge case
> - when the contour line hits the edge...
> 
> Barry
> 
> _______________________________________________ R-sig-Geo mailing
> list R-sig-Geo at r-project.org 
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
> 

- -- 
Edzer Pebesma
Institute for Geoinformatics (ifgi), University of Münster
Heisenbergstraße 2, 48149 Münster, Germany. Phone: +49 251
83 33081 http://ifgi.uni-muenster.de GPG key ID 0xAD3A77F1
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQEcBAEBAgAGBQJSVWfqAAoJEM1OCHCtOnfxDnsIAIapaJ3hiSN5D1Y7e4uUEYeH
rg8vvTCCzQjQxc1472wyUZYfK552kUugA+C0IYP97HWAEzMQkkPVI1EFCm2f3Cpo
6OBkSMvfaU+TM/Q7WMnDJqry0AKXXc576fgpAnYXktavgzzuEDbaFg5yAg717ZYz
SBrThQ8b0148DtWEHGsXQbBJvfxQY6ymhlPWBNQD2hKsqN+A/j27RADCdehmDnNk
ZkMUGN2ehxKrX9MrkdNU0JFLr41diKzNeqsULzr4YH2yyzTiQqKsPdF7PvmuEdqK
2tGNpOi2SWimLWcXpdXQ2NqGVhKkPzC5Pj0K4fXBs4n4QES2Q6cc2XgKJOeCaKU=
=qGv9
-----END PGP SIGNATURE-----