[gdal-dev] Why is the nearest neighbour resampling method the "worst interpolation"?

[Matt Wilkie]

Hi Vincent,

 > Would people on this list have some knowledge or maybe hints to
 > documents on what downsampling/interpolation method would be best for
 > elevation data, for example when downsampling from 90 to 250 or 500
 > meters pixel size, keeping in mind that the result will be used to
 > calculate slope/aspect from?

In my experience with elevation data Bilinear interpolation is the best 
to use when downsampling. Nearest Neighbor will often produce distinct 
box like artifacts or streaking, while Cubic Spline gives a kind of ripple.

With elevation it is important to keep the numbers to the right of the 
decimal for calculating slope/aspect (meaning: don't convert to integer) 
even when grossly downsampling. I can't locate the paper where I learned 
this at the moment, but it was in connection with ANUDEM, Mike 
Hutchinson and the Australia National Elevation Model.

cheers,

matt wilkie
--------------------------------------------
Geographic Information,
Information Management and Technology,
Yukon Department of Environment
10 Burns Road * Whitehorse, Yukon * Y1A 4Y9
867-667-8133 Tel * 867-393-7003 Fax
http://environmentyukon.gov.yk.ca/geomatics/
--------------------------------------------