<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
----- Message d'origine ----- <br>
De : Antoine Monmayrant <br>
Date : 14/10/2010 14:13:
<blockquote cite="mid:4CB6F3EF.8000100@laas.fr" type="cite"> Watch
out with this one! <br>
Contrary to Matlab griddata that does linear interpolation,
"cshep2d" uses cubic
<meta http-equiv="content-type" content="text/html;
charset=ISO-8859-1">
For scattered data, you must use cshep2d() and eval_cshep2d(). <br>
<span style="visibility: visible;" id="search">shepard that
introduces a lot of under- and over-shoot between sample data
points.<br>
It looks nicer than linear interpolated data, but can be a bit
misleading.<br>
</span></blockquote>
This is the interpolation world...<br>
AFAIK, there is no equivalent to griddata() in Scilab (core)<br>
<br>
According to its help page, linear_interpn() takes poles on a grid <br>
to interpolate on scattered points, but not the reverse, whereas<br>
original data from Nobuyuki Hosomi are scattered.<br>
<br>
Maybe a linear solution could be found in a Scilab external package
?<br>
<br>
Samuel<br>
<br>
</body>
</html>