I searched the forum and the knowledge base but couldn’t find an answer to my question. So, there it goes:
How do I import plain text data into blender?
In my case it’s a digital terrain model containing x-y-z values (x & y being the Gauß-Krueger coordinates in meters – similar to UTM –, z being the heights) that I want to import. Values are of float type and separated by tabs. I’d also have a dbf-file with this data, if it helps any further.
you’ll need to write a script and translate the coordinates into blender-vertices and connect them to form a mesh. you might wanna check out the python scripting documentation on blender.org. also post the question on the python forum here, this will catch the attention of more ppl which might be able to help
Is the DTM already a TIN (triangulated irregular network) or in grid format. If it isn’t (its just random spot elevations) you will have to do a delaunay triangualtion first which converts your data to triangles which can be given faces when you import. You may also want to convert your data to a regular grid which would be able to be subsurfed nicely. Writing a program to to the Delaunay triangulation isn’t difficult but its not trivial either. There is an open source GIS program called GRASS which would be able to read your data and create a TIN and/or grid and export to dxf and/or vmrl. You could then import the vmrl into blender.
GRASS by the way can also export 3d views and movies of the data so that Blender may not be required at all depending on your purpose. The web site for the GRASS project is here: http://grass.itc.it/
Let me put some background to my question. I’m involved in a (very little) project by my university which wants to make a 3d town model (standing pictures and animated flight) for a small town nearby.
We have quite a bit of data:
high resolution evaluation model (data form laserscanning with random positioned height data that distinguishes between points on the ground and on top of buildings),
aerial photos of the area (of good quality)
cadastral data containing the shapes of buildings and
phots of some house fronts
Now, there are different approaches towards it. One way would have been using commercial products like e.g. from CyberCity and ViewTec. But they cost a lot of money (some 10-thousand Euros). Then, we could get some licences of the LandExplorer by the HPI Potsdam, but possibilities concerning textures seem to be limited. Now, I had the idea to maybe integrate our data into blender, having better opportunities modelling man-made 3d objects (like houses, etc.).
I converted my data into a TIN and will export it to VRML (when I’m back at university next week), then it should work fine, importing the data into blender. Then I’ll see how to wrap the aerial photos and deal with the buildings.
Well, there will be probably arising more questions later on, but thanks for now! If anyone has more ideas about this …