Michael Bedward ha scritto: > On 22 March 2010 21:57, Andrea Aime wrote: >> Maybe by adding missing points outside of the specified range >> to make the cubic algorithm work? >> By the way I read the spec we can imagine repeating the first >> and the last point indefinitely below and above what was >> specified without really changing the outcome? >> So I guess we could add one copy of the first point below >> the first point and one copy of the last point above the >> last point just to handle the corner case of X falling >> in the area we cannot cover with the given points. >> Wondering, would doubling the point in place work? >> Like, we assume the first point is there two times in >> the expression? And same for the last? >> Would that provide a continuous output? > > Here's a test I did with the cubic algorithm... > http://imagebin.org/89895 > > The first and last points are added such that they like outside the > range of the "real" interpolation points but have the same y-value as > their neighbour. For the cubic algorithm it is actually a requirement > that no points have tied x values - I'll have to add a check for that > to the function. > > So there could be this scheme: > > N points Algorithm > 0 any throw an Exception > 1 any evaluate point's value and return for all inputs >> =2 lin / cos ok > 2 cubic fall back to linear (since 4 collinear > points would give ~linear results) > 3 cubic add dummy point before or after to put > lookup value into > middle interpolation interval >> =4 cubic ok > > Plus, for 2 or more points, treat lookup values outside the range of > the points as per the spec. > > What do you think ?
Sounds good to me Cheers Andrea -- Andrea Aime OpenGeo - http://opengeo.org Expert service straight from the developers. ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ Geotools-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/geotools-devel
