I have some delimited-text point data that is encoded as X,Y positions relative to a given coordinate. So, for example, 1 5 represents a point 1 meter east, five meters north of the origin coordinate. It is essentially a UTM projection but does not use a standard zone. What is the best way to import this data into QGIS?

  • Yes, you can figure out some sort of process in QGIS involving custom projections, or PyQGIS scripting. It involves a bit of research and thinking. But if it is only one file, you can just open it in Excel, apply a quick formula to convert to the coordinate of the origin point, and then open in QGIS - with little research or thinking. – jberrio Jan 22 '19 at 6:28

If you want to use a UTM-like projection that isn't a standard UTM zone, you can just specify your own Transverse Mercator coordinate system.

Here's the PROJ.4 page on Transverse Mercator projections.

If your grid origin is known in geographic Longitude, Latitude coordinates, your projection will look like:

+proj=tmerc +lat_0=ORIGIN_LAT +lon_0=ORIGIN_LON +k_0=1 +ellps=WGS84 +units=m +no_defs

To create a custom projection in QGIS, go to Settings > Custom Projections... > [+] sign and then give it a name and supply the projection definition from above.

Then to load a CSV, go to Layer > Add Layer > Add Delimited Text Layer and click the globe button next to Geometry CRS. Search for your custom projection, then confirm until your data is added to the map.

  • Adjust the false easting and northing by your known coordinate and this homebrew CRS becomes a local datum, used often in mine sites and by users of CAD software... because it keeps the coordinate values small (for example instead of 512341, 7123455 using a local datum becomes 1,5 - less chance of misreading). – Michael Stimson Jan 22 '19 at 0:06
  • 1
    I used the rule you suggested and it seemed to work, except that my point cloud is significantly offset from where it ought to be. This section of the FAQ seems to be relevant to what's happening, but adding +nadgrids=@null to my specification does not solve the problem. I think perhaps +ellps=WGS84 is not the ideal ellipsoid but I don't know what option would be better. The area I'm working on is in Norway, so it's pretty far north, if that helps. – John Allie Jan 22 '19 at 18:37
  • Here is an image of the point cloud overlaid with other (correct) chart shapes: imgur.com/a/wEH4y8h – John Allie Jan 22 '19 at 18:41
  • @JohnAllie It indeed does like a datum/ellipsoid difference. Maybe you can view the CRS information for your polygon file and see what datum/ellipsoid it is in. – Logan Byers Jan 22 '19 at 20:04

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.