Exporting a DTM (TIN) to ASCII Gid Format (ESRI - ArcGIS)

[Paul_Ollett]

Hi there,

Is it possible to export a DTM (TIN surface) to ESRI's ASCII grid format (for ArcGIS)?

I've attached an example file.

Failing that, is it possible to export to a simple XYZ point file?

Cheers,

Paul.

[Neilw_05]

That's a shame. ArcGIS can convert the Geotiff to ASCII or grid. It would be a one step process to export to GeoTiff.

I can imagine how many hours you have spent trying to resolve this, all to save someone the bother of pushing a button.

[ouspensky]

Here is my follow up to an earlier post about using FOSS (Free and Open Source Software) GIS to sample raster grids off of C3D TINs:

The raster sampling in C3D outputs corrupted files that cannot be read in addition the only sampling option is NEAREST NEIGHBOR INTERPOLATION which also sucks and outputs grids with all kinds of stepped artifacts that are generally unsuitable for hydrology software.

A much better solution is the output the points to GRASS GIS and use the various interpolation options to generate a grid of any resolution, angle, and quality then output to any OGR format (which includes ESRI Grid format, either ASCII or Binary).

The steps are export a CSV report from C3D:

then set up a GRASS location to match the drawing coordinate system:

 

then define the computation extents for interpolation called "Region" in GRASS

here is where you can set the grid size for sampling

Then import the points using V.IN.ASCII

Then interpolate the surface from points raster > surface > v.rst.surf or whichever you want

Finally Export to ESRI Grid or geoTIFF or whatever:

I included a dataset

[ouspensky]

I wanted to follow up the post with an attachment of screenshots used in the process

[antoniovinci]

---

ouspensky wrote:

The steps are export a CSV report from C3D:

---

Well, in your post 09-02-2014 04:47 AM you wrote something else:

"To use the QGIS method you can import the LandXML then extract the points using the QGIS vector tools"

while your tutorial does NOT involve any LandXML, but only a simple points' export...

By the way, to make a Dem out of a point list, you can use only Qgis, without Grass.

[neilyj666]

It is also possible to use kriging to smooth Civil 3D surfaces and extrapolate beyond known data - never used it so don't know how useful/efficient etc it is....

There should be more GIS type interpolations available e.g. Inverse Distance Weighting but who knows why they are not included...? I've just downloaded QGIS to see what it can offer.

[ouspensky]

In theory QGIS is a front end for GRASS, but I included a QGIS workflow in
the dataset attachment. I like to go straight to GRASS these days for the
raster work. The purpose of using GRASS is to control the way the points
are interpolated as much as file conversion. I thought that CSV points ->
grid was better than XML because it strips out all the headers and tags,
smaller, faster, editable in excel, ect.

[Neilw_05]

One thing that needs to be considered in multistep workflows is that each step may require a considerable amount of processing time for large datasets. More steps equals more processing time. Unless there is a specific need for the extra steps it would be counter productive to take that route.

[ouspensky]

yes I agree. to me it is a three step process.

1. export CSV from c3d

2. import CSV into GRASS

3. export DEM from GRASS

GRASS is an extraordinarly stable software, more so than QGIS, more so than anything from autodesk.  It can handle extremely large datasets.

the extra step is because c3d only offers nearest neighbor interpolation of points and this will produce artifacts and "steps" along contours. in addition there is something corrupt in the files that it produces.

by contrast GRASS allows you to interpolate by regularized spline with tension or inverse weighted distance or you can go off the deep end with geostatistics and kriging if necessary using Postgre POSTGIS and R statistics package (also 100% free) whatever you need to get a DEM that meets the project demand for analysis in hyrdology software.

c3d is a great vector software but it simply has no capacity for generating actual usable raster grids

the only thing that I have ever found that plugs the hole is GRASS or arcGIS

since you can script Python directly into GRASS, steps 2 and 3 from above can become a one button push kind of thing...

[Neilw_05]

I know about the defect when exporting to DEM format as I mentioned in an earlier post to this thread. Support says all we need to do is increase the sampling density to a 1 foot grid. Like that is going to fly with a huge DEM. I guess we won't be seeing a fix anytime soon with that "solution".

I've not seen the artifacts you mention, but I haven't yet worked with an exported dataset thanks to the DEM defect. I did find the GeoTiff produces a useable Raster but again I haven't worked with the results enough to notice any problems.

A small grid interval should address any stepping problem but that may not be practical for a large dataset.

How do GRASS processing times compare to C3D?

[Neilw_05]

I have a task that is related to this post which I need to solve. I have a very large LIDAR dataset consisting of an ASCII points file and some 3D polylines in a CAD file that I have used to build a surface. My surface encompasses only a small area within the dataset so I have applied a clip boundary to limit the data read from the points file. Now I want to export the points used in my surface to an ASCII file so that other applications can utilize the points without having to extract them again from the original file.

Is there a way to export the surface points to a CSV? I know I could display my surface points, explode the surface and export the points, but that will likely take a while to process all the steps, so I am looking for a more direct solution.

[sboon]

One of the reporting tools produces a csv file.  It's slow, and you have to remove the extra lines it creates but it will work - eventually.

Steve
Please use the Accept as Solution or Kudo buttons when appropriate