I'm trying to figure out how I can preserve my GCP's after a gdalwarp. I need the georeferenced GCP's data, so I can calculate the Root Mean Squared Error.

Is there a way to preserve this data, and somehow use it, or are there other ways to calculate the RMSE of a georeferenced raster file?

  • Have you tried saving the warping result as .vrt? This should give you an editable file. – AndreJ Jun 16 '14 at 4:55
  • I'm currently warping to a GTiff, I'll give the .vrt a try! – Danny Hoek Jun 16 '14 at 9:32
  • You could warp to both formats in a batch if you prefer to work on with a Geotif. – AndreJ Jun 16 '14 at 10:58
  • 1
    You should not really calculate RMSE by using the same GCPs which were used for computing the warping. With 3 GCPs and 1st order polynomial transform the RMSE will be zero, as well as always with thin plate spline transform. The result does sometimes reveal bad GCPs. For reliable results measure new independent check points. – user30184 Jun 16 '14 at 12:52
  • 1
    When people are doin quality checks before paying to orthophoto contractors they use independent checkpoints which are collected with better accuracy than what is the demand for the end product. There are lots of litterature about the methods, first to meet the eye was asprs.org/a/publications/pers/2001journal/february/…. – user30184 Jun 16 '14 at 16:02

Your Answer

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

Browse other questions tagged or ask your own question.