Effacer les filtres
Effacer les filtres

different results with griddata in Matlab2011b vs Matlab2008a

1 vue (au cours des 30 derniers jours)
DF
DF le 9 Fév 2013
I am using the griddata function with exactly the same variables in Matlab2008a and Matlab2011b, but I get different results. The result from 2011b does not seem correct to me.
The command I'm using is:
[V1 V2 V3] = griddata(A(:,1),A(:,2),A(:,3),B(:,1),B(:,2));
And I check the number of NANs:
S=sum(isnan(V3))
In Matlab2008a I get 6 NANs, in Matlab2011b I get 20.
The A and B matrices I use can be downloaded from here: A_B_matrices
(I give you the exact matrices because I can not reproduce the problem with any numbers. However, they are small: 21x3 and 27x3)
Can someone tell me what has changed between the two Matlab versions I'm using and how can I get the same results in both of them?

Réponses (1)

Shashank Prasanna
Shashank Prasanna le 9 Fév 2013
Take a look at the release notes. There have been few changes in:
R2009b > Mathematics > Computational Geometry Functions Being Changed
2- and 3-D computational geometry functions (delaunay, convhull, griddata, voronoi, delaunay3, griddata3) no longer use QHULL or the QHULL options arguments. The N-D functions gridatan, delaunayn, convhulln, and voronoin still use QHULL. Compatibility Considerations
The QHULL options arguments to delaunay, convhull, griddata, and voronoi are no longer required and are currently ignored. Support for these options will be removed in a future release.
  1 commentaire
DF
DF le 11 Fév 2013
Yes, but I don't use any QHULL options when calling the griddata. So, I don't think this is the root cause.

Connectez-vous pour commenter.

Catégories

En savoir plus sur Delaunay Triangulation dans Help Center et File Exchange

Tags

Produits

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by