0

In connection with an archaeological project, I have data in .dxf format (lines and polygons). When I import them to QGIS, however, it positions them 8000 km southeast. I tried to change the CRS (WGS 84; WGS 84/UTM zone 35) of the files, but without success. The input data are in the local coordinate system 1970 I guess. I also tried to save .dxf file as .shp in different CRS - no effect.

Is there a way to fix the problem?

PolyGeo
  • 65,136
  • 29
  • 109
  • 338
Plamen Georgiev
  • 21
  • 5
  • 1
    Welcome to GIS.SE. We need some more information. How were the .dxf created? If it was done using CAD and whoever drew this didn't bother to emulate a CRS and simply created the features around 0,0, you'll have to take it back to them and ask them to move the data to the correct place. Also, which CRS does QGIS initially assign to the .dxf? How exactly did you try to change it? Why are you guessing, which CRS the data is/should be in? Please edit your question accordingly. – Erik Sep 30 '20 at 07:31
  • 1
    It sounds like a local datum which is common for mine sites, building plans and archaeological excavation areas. If it is a local datum and the site is fairly small the scaling of the features should be OK it's just the location that's out, you can generate a custom CRS with an adjusted false easting and northing https://gis.stackexchange.com/questions/203262/whats-the-best-way-to-handle-a-site-datum-in-qgis should give some idea of your next step, you may need to affine the data if the scale/rotation is still wrong compared to known accurate data. – Michael Stimson Sep 30 '20 at 07:42
  • 2
    https://gis.stackexchange.com/questions/33208/georeferencing-vector-layer-with-control-points-using-qgis will help if the data is not scaled/rotated correctly, you will need to find at least 3 points of similarity to affine and preferably more than 3 to get some sort of error correction, preferably spread out over the site and at the ephemera where rotation around the central point is most pronounced. – Michael Stimson Sep 30 '20 at 07:45
  • You mentioned 1970. Is the data in Romania and you think they may be in Stereo 70? Try defining one as 3844 (Pulkovo 1942(58) / Stereo 70 instead. – mkennedy Sep 30 '20 at 22:58
  • As alternative to de standard cad import you could try this qgis plugin https://plugins.qgis.org/plugins/AnotherDXF2Shape/ – Hans Erren Nov 10 '20 at 16:37

1 Answers1

2

The plugin AnotherDXF2Shape works fine, but was too late. I asked the surveyors for a list of points in .txt-file. I transformed the coordinates from the local system into a world file and imported them into QGIS 3 as CSV. Then I connected the points in lines and polygons and saved the files as shapes. I lost some of data, but it was not crucial to me.

Kadir Şahbaz
  • 76,800
  • 56
  • 247
  • 389
Plamen Georgiev
  • 21
  • 5