1
<?xml version="1.0" encoding="ISO-8859-1" standalone="no"?>
2
<!DOCTYPE MAN SYSTEM "../../manrev.dtd">
4
<LANGUAGE>eng</LANGUAGE>
5
<TITLE>eval_cshep2d</TITLE>
6
<TYPE>Scilab Function</TYPE>
7
<DATE>February 2004</DATE>
8
<SHORT_DESCRIPTION name="eval_cshep2d">bidimensional cubic shepard interpolation evaluation</SHORT_DESCRIPTION>
11
<CALLING_SEQUENCE_ITEM>[zp [,dzpdx, dzpdy [,d2zpdxx,d2zpdxy,d2zpdyy]]] = eval_cshep2d(xp, yp, tl_coef)</CALLING_SEQUENCE_ITEM>
17
<PARAM_NAME>xp, yp</PARAM_NAME>
19
<SP>: two real vectors (or matrices) of the same size</SP>
23
<PARAM_NAME>tl_coef</PARAM_NAME>
25
<SP>: a tlist scilab structure (of type cshep2d) defining a cubic Shepard interpolation function
26
(named <VERB>S</VERB> in the following)</SP>
30
<PARAM_NAME>zp</PARAM_NAME>
32
<SP>: vector (or matrix) of the same size than <VERB>xp</VERB> and <VERB>yp</VERB>,
33
evaluation of the interpolant <VERB>S</VERB> at these points</SP>
37
<PARAM_NAME>dzpdx,dzpdy</PARAM_NAME>
39
<SP>: vectors (or matrices) of the same size than <VERB>xp</VERB> and <VERB>yp</VERB>, evaluation
40
of the first derivatives of <VERB>S</VERB> at these points</SP>
44
<PARAM_NAME>d2zpdxx,d2zpdxy,d2zpdyy</PARAM_NAME>
46
<SP>: vectors (or matrices) of the same size than <VERB>xp</VERB> and <VERB>yp</VERB>, evaluation of the
47
second derivatives of <VERB>S</VERB> at these points</SP>
55
This is the evaluation routine for cubic Shepard interpolation function computed with
56
<LINK>cshep2d</LINK>, that is :
59
zp(i) = S(xp(i),yp(i))
60
dzpdx(i) = dS/dx(xp(i),yp(i))
61
dzpdy(i) = dS/dy(xp(i),yp(i))
62
d2zpdxx(i) = d2S/dx2(xp(i),yp(i))
63
d2zpdxy(i) = d2S/dxdy(xp(i),yp(i))
64
d2zpdyy(i) = d2S/dy2(xp(i),yp(i))
68
<SECTION label="Remark">
70
The interpolant <EM>S</EM> is C2 (twice continuously differentiable) but is also extended by
71
zero for <EM>(x,y)</EM> far enough the interpolation points. This leads to a discontinuity
72
in a region far outside the interpolation points, and so, is not cumbersome in practice (in a
73
general manner, evaluation outside interpolation points (i.e. extrapolation) leads to very
79
// see example section of cshep2d
81
// this example shows the behavior far from the interpolation points ...
82
deff("z=f(x,y)","z = 1+ 50*(x.*(1-x).*y.*(1-y)).^2")
85
X = X(:); Y = Y(:); Z = f(X,Y);
87
// evaluation inside and outside the square [0,1]x[0,1]
89
xx = linspace(-1.5,0.5,m);
90
[xp,yp] = ndgrid(xx,xx);
91
zp = eval_cshep2d(xp,yp,S);
92
// compute facet (to draw one color for extrapolation region
93
// and another one for the interpolation region)
94
[xf,yf,zf] = genfac3d(xx,xx,zp);
95
color = 2*ones(1,size(zf,2));
96
// indices corresponding to facet in the interpolation region
97
ind=find( mean(xf,"r")>0 & mean(xf,"r")<1 & mean(yf,"r")>0 & mean(yf,"r")<1 );
100
plot3d(xf,yf,list(zf,color), flag=[2 6 4])
101
legends(["extrapolation region","interpolation region"],[2 3],1)
111
<AUTHORS_ITEM> Robert J. Renka</AUTHORS_ITEM>
112
<AUTHORS_ITEM> B. Pincon (scilab interface)</AUTHORS_ITEM>