This toolbox contains code for 2D multivariate interpolation in MATLAB.
- Adaptive Normalised Convolution (ANC)
- Radial Basis Function Interpolation (RBF)
- Kriging
- Natural Neighbour Interpolation
- The included Natural Neighbours interpolation software was provided by Pavel Sakov and is available at URL: http://www.marine.csiro.au/~sakov/nn.tar.gz
Unless otherwise stated, all code is Copyright Matt Foster ee1mpf@bath.ac.uk
Some functions require MEX functions to built. To compile all of the necessary
Makefiles in UNIX, type make
. Some of the Makefiles may require alterations
for your system, however, provided you have a complete build environment, and
the binaries mex
and mexext
are available, the process should be fairly
simple. You should then find a directory named toolkit
. Copy this directory somewhere in your MATLAB work area, and finally, from within MATLAB, issue the command:
addpath('interpolation_toolkit');
It is not necessary to use genpath
, or include the private
directory
With the exception of ANC, which interpolates onto matrices, all of the functions provided have prototypes of the following form:
[xx, yy, zz] = function(xi, yi, zi, xx, yy, [optional args])
or more simply
zz = function(xi, yi, zi, xx, yy, [optional args])
The help within the various functions gives more information on the input and output arguments. For example, issuing the command:
help rbf
Will print the built in help for the RBF interpolation command.
Running the command:
help interpolation_toolkit
Where interpolation_toolkit
is the directory name, will give basic help for the toolkit, including some usage instructions.
ANC operates using matrices, and requires arguments of the following form:
out = adaptiveNC(si, cm, [optional args])
Full help is available within the m-file.
Please report bugs to Matt Foster ee1mpf@bath.ac.uk
Moved from local SVN to GitHub: 20090120