Fwd: Interpolation on scattered data

[Nicklas Karlsson]

Den mån 23 mars 2020 kl 18:30 skrev c. <address@hidden>:

> Il giorno 23 mar 2020, alle ore 18:09, Nicholas Jankowski <address@hidden> ha scritto:
>
> It's not possible to send your own triangulation to griddata(...) and
> I need to do because griddata(...) function does not work, it actually
> crash Octave. Data come from measurement on many different point in
> system and hopefully same triangulation work for interpolation for all
> values.
>>
>> Think the best option is to modify griddata(...) function to my need. It only took a minute or so to find the griddata function on filesystem, it is a rather short function. Need to supply my own triangulation and it is also useful to show triangulation used so to check it is reasonable.
>>
>> Only drawback modifying griddata(...) function is I have to make it available to others to fulfill license requirements. ;)
>>
>
> (FYI please keep the mailing list in the conversation, and it is the
> convention of this mailing list is to bottom post.  see
> http://www.idallen.com/topposting.html )
>
> can you provide a sample of the data that is crashing Octave? I'd be
> curious to see what the specific problem is.
>
> Also, I'm no expert on licensing, but my understanding is you can make
> and keep private any changes to the code you want unless you
> distribute your modified program in some fashion, in which case you
> need to make the source code available along with it.
>
> That said, if it improves the function while maintaining compatibility
> I'm sure we'd love to consider any improvement to the code.
>


I actually already have a modified version which I used locally.
I attach it in case it is of use to anyone else.
c.

Tried it and it works!

griddata(...) function work on gridded data while this function work on triangulation. Will write a few a words about it tonight and suggest it is added to Octave among the "tri" functions as it belong there.

 

There already are triplot(...) trimesh(...) trisurf(...) and suggest tridata(...) would be a good name for this function. It is probably a good idea to use similar parameters as on the other tri functions tridata(tri, x, y, xi, yi) but are a little bit unsure about method if higher degree was ever to be added, know lagrange polynoms are commonly used for partial differential equations.

Thanks a lot, Nicklas Karlsson

PTC2griddata.m
Description: Text Data