There is no chance for an algorithm to know what points to use for the linear 
interpolation interval. Without knowing the properties of the terrain, knowing 
that there is a valley floor, that point A and B are good ones and that the 
terrain can be represented by a line, it's impossible to define the intervals 
used for interpolation.

In such a case you have to draw a very coarse track with just a few points at 
very well defined points. This will give you a rough estimate about the ascent 
and descent. The distance will be wrong of course.

The core problem is the bad DEM data. It lacks the information needed. There 
is no way to create better information out of such data without having another 
source of information to correlate with. 



Am Montag, 24. Juli 2017, 21:14:43 CEST schrieb Guido Scholz:
> Am Mon, 24. Jul 2017 um 20:30:08 +0200 schrieb Oliver Eichler:
> > Hi Guido,
> 
> Hi Oliver,
> 
> > I do not understand why the elevation filter "Interpolate Elevation Data"
> > with setting "coars" does not fit your need. The used spline
> > interpolation is much more natural than a linear interpolation.
> 
> yes and no. The spline interpolation provides no chance for improvement
> for very narrow valley situations like in the attached graph. The plot
> around 15 km +-3 km in reality is at 2150 m (Río Marañon valley). A
> linear interpolation would give a chance to eliminate this error (more
> than 200 m!) completely.
> 
> > But either linear or spline interpolation will not provide good results if
> > your input data is faulty. Especially along steep walls a few meters
> > deviation in the horizontal plain can result in a significant difference
> > of the elevation. In that case you have to reduce points to a minimum.
> > And fine tune the position to a better fit.
> 
> Yes that is clear but this is no solution for narrow valleys where DEM
> data gets wrong due to its "natural" behavior.
> 
> The linear interpolation would just be a tool to make elevation gain
> calculation much easier. It would also help in your case of a not so
> precisely painted route.
> 
> Guido



------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Qlandkartegt-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/qlandkartegt-users

Reply via email to