5

I am trying to transform from epsg:4326 (wgs84) to epsg:25832 (utm ETRS89) using Proj.4. I am comparing the results with results from http://valdemar.kms.dk/trf/ .

it seems like the results calculated by proj.4 is off by 2m northing and 4m in easting in Copenhagen. The results farther west are more correct.

Am i missing some sort of correction attributes in my calculation?

|improve this question
1

I've tested this using proj.4 (4.7.1) and can't reproduce this:

Using coordinates from a point just west of Kastellet in Copenhagen the discrepancies are within centimeters:

Using Valdemar:

from geo_etrs89 -> utm32_etrs89
725627.5m  6177576.0 m

And using cs2cs: 

\FWTools2.4.7>cs2cs +init=epsg:4326
+to +init=epsg:25832
12.59 55.6914
725627.46       6177576.01 0.00

Could the problem lie in the parameters you used with proj.4?

The discussion here may also share some light to the questions: http://trac.osgeo.org/proj/ticket/15

|improve this answer
  • What is the proj4 strings that you use for the transformation? – Jesper Kihlberg Oct 23 '10 at 17:33
  • +init=epsg:4326 +to +init=epsg:25832 - the parameter string functions with both proj and cs2cs. – haakon_d Oct 24 '10 at 12:30
0

Neat question - lead me to investigate ETRS89 a little. Interesting bit of information (for me, anyway) is from this explanation of the derivation of ETRS89. It does make me wonder if one of the methods you're using is accounting for drift, and the other isn't. Seems like a lot of difference though, so probably not the entirety of the problem.

I noticed that http://valdemar.kms.dk/trf/ doesn't have an input for WGS84; what did you use for the source system?

|improve this answer
  • I am using the input "geo_etrs89", as etrs89 is the european version of wgs84. wgs84 and etrs89 should be the same, as far as I have understood.. – Jesper Kihlberg Oct 13 '10 at 13:58
  • Well, they're really similar; etrs89 is based on wgs84 but accounts for continental drift, apparently (see link 1, above). If you used geographic etrs89 for the input to both methods, I'd have thought you'd have gotten very nearly identical results. – Herb Oct 13 '10 at 14:19
  • I tried to use etrs89 as input for both, and the results are exactly the same, still off by 2m and 4m.. – Jesper Kihlberg Oct 14 '10 at 19:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.