I'm working on georeferencing old maps for archaeological purpose. Of course, colonial maps all do have some degree of transformation needed. But sometimes a specific region is well mapped while the neighbouring one is not. So I wonder if there is a way to calculate the amount of (in)accuracy, or the level to which the original image has been transformed in order to fit the actual landscape? Could a new raster be generated with these values? I'm using QGIS and TPS for georeferencing, but it can be on any platform.

  • 1
    as far as i remember, the accuracy is normally calculated from the ground control points, not the image itself. i dont know what qgis plugin you are using, but it seems the docs.qgis.org/2.8/en/docs/user_manual/plugins/… includes the residual as well as the mean error for the GCPs (figure 1) – Jonas Aug 21 '18 at 11:22
  • So the overall accuracy is expressed by the global mean error, and each GCP has its own residual. I could compare two maps with the same control points, and see which one is more or less accurate. But how could I use these values to analyze a single map? Split it into tiles maybe? – gvanhavre Aug 23 '18 at 9:51

I'm not sure if this is what you want but when you load a map with 4 or more calibration points in MAPC2MAPC, it calculates a best-fit affine transformation between coordinates and pixels - and then reports on how good the fit is.

If you have an image and a suitable calibration file it will only take seconds.

But what to do if the calibration is 'off' is another matter!


  • 1
    You should add that you are the author of the program. – AndreJ Sep 23 '18 at 11:40

Load the maps shapefile from a trusted source and compare them. If there is some error it can be further improved by taking more control points, changing transformation parameters and trying a different datum.

  • @Damini Jain, I think it is not that easy to find archaeological shapefiles from a trusted source. I assume it could more efficient to find archaeological web map service (WMS, TMS/WMTS etc.) and use it as a background map. – Taras Aug 21 '18 at 4:49
  • There are many sites of government and sources from where you can get it. You have to look on the Internet. So my answer isn't wrong. – Damini Jain Aug 21 '18 at 6:00
  • @DaminiJain, thanks, but georeferencing itself is not the problem. I'm looking for a way to evaluate the degree of transformation of the original image, like on a raster calculation. If it matches archaeological sites or not will be a question for later. – gvanhavre Aug 21 '18 at 9:13
  • @gvanhavre geog.uoregon.edu/amarcus/geog418/Labs/Lab06_rectification.htm hope this helps. – Damini Jain Aug 21 '18 at 10:15

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.