1 reply to this topic

[Martin Gamache]

We've noticed a potentially disturbing shift when resampling a 25m BC TRIM DEM down to 90m to match SRTM data. I'm resampling a full DEM of BC at 25m and the resulting shifts which we have not quantified do not seem to be systematic across the province but seem to change direction and magnitude. We have tried resampling using all three interpolation methods and were expecting a shift with nearest neighbor but were surprised to also find a shift with CC and BILinear.

Does, anyone have any clue what might be causing this shift? Thinking the pixel size to be part of the problem we have tried resampling to 100m and 75m and have observed the same issue. The shift also seems present in at least one other piece of software (PCI Focus) so we are not sure if it is caused by the way these two pieces of software specify the pixel origin (this would imply a systematic shift however) nor have we closely examine whether the shift varies between the two software.


mg

[GISRox]

mg- I've read this post a couple times. I might be able to answer the question, but I'm not sure if your problem is with the resampling algorithms or in the geo-referencing. Can you provide some sort of graphic to show the problem?