I'm a newbie where GIS is concerned, so please don't hesitate to point me to a manual.

We've got DEM files without any projection information. One is called 000426_LO17_200M_BD16_25.ORT and contains lines like the following:

-75925.00 2710900.00   834.88
-75925.00 2710800.00   832.98
-75925.00 2710700.00   830.77
-75925.00 2710600.00   832.38
-75925.00 2710500.00   830.83

So we think this is a LO17 projection, at a 200M DEM spacing, and we're guessing the 3rd column is elevation in m. Given that we know the data is from South Africa (or close to it), this seems about right.

We're trying to get this data into Google Fusion, but need to convert it into a format that Google Fusion will understand, or equivalently, into a gdal format that we can then convert to Google Fusion.

We've tried a gdal_grid conversion, but when attempting to import into Fusion, we get Fusion is complaining about no mapping for TMSO to proj.4

If anybody has any ideas or suggestions, it would be greatly appreciated.

All the best, Craig

  • 1
    can you try gdalinfo for your grid for getting whether projection is correct or not?
    – urcm
    May 22, 2012 at 11:59
  • 2
    From the name I'd expect the same as you suggest except I notice that what I assume is the y coordinate is not incrementing in 200m intervals, but only 100m intervals. I agree that the third colum is most likely to be elevation. Can you provide a link to one of the files, as it will be easier to understand if we can see the data itself. May 22, 2012 at 20:47
  • 1
    Hi Craig, are the five points that you provided above supposed to plot on the Hardap National Park, about 5 km west of the reservoir?
    – nhopton
    May 23, 2012 at 10:27
  • 1
    You need to determine the correct SRID, which is either EPSG:22277 or EPSG:2047
    – Mike T
    Jul 23, 2012 at 0:37

2 Answers 2


You can convert the file to Lat Lon Alt using:

invproj -f %.9f -s -m -1 +proj=tmerc +ellps=WGS84 +datum=WGS84 +k_0=1 +lon_0=17 +lat_0=0 +x_0=0 +y_0=0 +units=m <000426_LO17_200M_BD16_25.ORT >000426_LO17_200M_BD16_25.csv

ogr2ogr can read the out csv.


it's difficult just having a few points and not knowing anything else. To be honest, I'd be very surprised if the following is the correct solution to the problem.

What I did was to make a csv file from your five points and load it into the most recent version of QGIS (1.8). I then made a custom CRS using the following string:

+proj=tmerc +lat_0=0 +lon_0=17 +k=1 +x_0=0 +y_0=0 +axis=wsu +ellps=WGS84 +towgs84=0,0,0,0,0,0,0 +units=m +no_defs

and set the layer CRS to this. I then exported the csv layer to a KML file to see where the points plotted onto the real world. But I really doubt that this is going to help you much, I'm assuming that your 'z' values are elevations and these are nowhere near right for where the points plot. Of course, had this method worked it would have provided a simple means of transforming your data, sorry. LO17 is such a peculiar projection, in fact I'm far from sure how QGIS handles the "+axis=wsu" in the parameter string.


  • nhopton: I work with Craig above. The data we have covers most/all of South Africa and hence there is a good chance there is overlap into Namibia in the NW corner which the above file extract might be. Can you elaborate on how you got to that conclusion? A toolchain description would be very helpful if you have one! Thanks Mike
    – user7722
    May 23, 2012 at 15:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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