bidimensional cubic shepard (scattered) interpolation
tl_coef = cshep2d(xyz)
a n x 3 matrix of the (no gridded) interpolation points (the i th row given the (x,y) coordinates then the altitude z of the i th interpolation point)
a tlist scilab structure (of type cshep2d)
This function is useful to define a 2d interpolation function when
the interpolation points are not on a grid (you may use it in this case
but splin2d is better for that purpose).
The interpolant is a cubic shepard one and is a C2 (twice continuously
differentiable) bivariate function s(x,y) such that :
s(xi,yi)=zi for all i=1,..,n
((xi,yi,zi) being the i th row of
The evaluation of s at some points must be done by the eval_cshep2d function.
The function works if n>= 10, if the nodes are not all colinears (i.e. the (x,y) coordinates of the interpolation points are not on the same straight line), and if there is no duplicate nodes (i.e. 2 or more interpolation points with the same (x,y) coordinates). An error is issued if these conditions are not respected.
// interpolation of cos(x)cos(y) with randomly chosen interpolation points n = 150; // nb of interpolation points xy = grand(n,2,"unf",0,2*%pi); z = cos(xy(:,1)).*cos(xy(:,2)); xyz = [xy z]; tl_coef = cshep2d(xyz); // evaluation on a grid m = 30; xx = linspace(0,2*%pi,m); [X,Y] = ndgrid(xx,xx); Z = eval_cshep2d(X,Y, tl_coef); clf() plot3d(xx,xx,Z,flag=[2 6 4]) param3d1(xy(:,1),xy(:,2),list(z,-9), flag=[0 0]) xtitle("Cubic Shepard Interpolation of cos(x)cos(y) with randomly chosen interpolation points") legends("interpolation points",-9,1) show_window()
- splin2d — bicubic spline gridded 2d interpolation
- eval_cshep2d — bidimensional cubic shepard interpolation evaluation
|5.4.0||previously, imaginary part of input arguments were implicitly ignored.|
|Report an issue|
|<< bsplin3val||Interpolation||eval_cshep2d >>|