[R] Voronoi-Diagrams in R

Raphael Päbst raphael.paebst at gmail.com
Wed May 21 13:34:03 CEST 2014


I believe you are right. A night of sleep has done wonders for my
understanding of the problem.

Thank you for your patience and help!

Raphael

On 5/21/14, Rolf Turner <r.turner at auckland.ac.nz> wrote:
>
>
> It sounds to me as though you are simply getting yourself flummoxed by
> the fact that the different packages produce their output in different
> formats.  The information in the output will be the same (as Boris has
> indicated) --- it will just be arranged differently.  Learn to interpret
> the output of each function (RTFM!) and you will find that there is no
> problem.
>
> cheers,
>
> Rolf Turner
> (author of "deldir")
>
> On 21/05/14 09:33, Raphael Päbst wrote:
>> Thanks for the answer!
>>
>> I'll post a sample tomorrow, I have however found the following:
>> triang.list() gives me the coordinates of the triangle's vertices,
>> while delaunayn() gives me the indices of those coordinates. Thus it
>> should be more or less simple to convert the output of deldir() into
>> that of delaunayn()
>>
>> Once again, thanks!
>>
>> Raphael
>>
>> On 5/20/14, Boris Steipe <boris.steipe at utoronto.ca> wrote:
>>> deldir() uses Lee and Schacter's algorithm, while the geometry package is
>>> a
>>> (partial) implementation of Barber et. al's Quickhull. Since both
>>> algorithms
>>> are correct, they should give the same results for the same data.
>>>
>>> How about you post a small input dataset and list the output that you
>>> need...
>>>
>>>
>>> B.
>>>
>>>
>>>
>>>
>>> On 2014-05-20, at 2:04 PM, Raphael Päbst wrote:
>>>
>>>> Thank you very much, this looks promising.
>>>>
>>>> I have a follow-up question however, probably due to my thickness when
>>>> it comes to the underlying math.
>>>>
>>>> I am translating (as closely as possible) some code that has
>>>> originally been written for Mathlab and uses the delaunay() function
>>>> there. Now, if I understand it correctly, Mathlab uses the same
>>>> Qhull-Library as a basis, as does the Geometry-Package. So in theory
>>>> delaunay(x) in Mathlab and delaunay(x) in R should give me the same
>>>> results for the same x, where x is a matrix of coordinates.
>>>>
>>>> If I use deldir(x1, x2) where x1 and x2 are the first and second
>>>> column of that same matrix above, is there any chance to get the same
>>>> result as with delaunayn() or at least something similar in format to
>>>> the result of delaunayn()?
>>>>
>>>> Many thanks again and please excuse my crude english.
>>>>
>>>> Raphael
>>>> On 5/14/14, Boris Steipe <boris.steipe at utoronto.ca> wrote:
>>>>> Try:
>>>>>
>>>>> install.packages("deldir")
>>>>> library(deldir)
>>>>> ?deldir
>>>>> set.seed(16180)
>>>>> x <- runif(20); y <- runif(20); window <- c(0,1,0,1)
>>>>> tess <- deldir(x, y, rw = window)
>>>>> plot.deldir(tess, wpoints="real", wlines="tess")
>>>>>
>>>>> Cheers,
>>>>> Boris
>>>>>
>>>>>
>>>>> On 2014-05-14, at 8:18 AM, Raphael Päbst wrote:
>>>>>
>>>>>> Hello everyone!
>>>>>> I have returned to R after a longish break and am currently working
>>>>>> on
>>>>>> a project where I need Delaunay-Triangulations and Voronoi-Diagrams.
>>>>>> If I understood it correctly, the Geometry-Package only offers
>>>>>> functions for the Delaunay-Triangulation at the moment. Is this
>>>>>> correct and if so, what would be the best way to get Voronoi-Diagrams
>>>>>> as well?
>
>



More information about the R-help mailing list