I'm trying to use photogrammetry and Structure from Motion to record buildings, and I'm looking for a way to bring these into a format usable in GIS, in particular pointclouds and models of terrain, buildings, and objects.

I would like to be able to import the pointcloud, model or rectified photographs into QGIS and manipulate it in terms of georeferencing and fixing axis (for creating rectified, scaled views of vertical walls/ objects). The software I'm using is QGIS, Agisoft Photoscan and Meshlab.

Does anybody have any suggestions regarding file formats, alternative (preferably opensource) software and workflow, or whether GIS is the proper format to do this in?

  • Ciao in my project (we are archaeologist and we are working on many ancient walls and structures) we are also looking to a solution for the same issue. Qgis is not the solution since it is not managing directly 3D features. We will give GRASS a try but we are not working on it before summer. I hope somebody can give us some ggod tips, too
    – seba
    Commented May 28, 2013 at 12:37
  • Thanks Seba. I'm also an archaeologist. I guess I'm looking for a fix - a way of capturing a particular perspective and keep scale, brought from Meshlab into a format usable in GIS - which will allow me to digitise. Perhaps just capturing the image (save snapshot) and importing it as a raster into GIS is the only way. But I thought it best to check in case there is a more elegant way to do this. Best of luck with your project! Richard
    – Richard
    Commented May 28, 2013 at 12:52
  • Take a look here, these guys are really good (all Open Source based): archeos.eu/wiki/doku.php?id=screenshots
    – markusN
    Commented May 28, 2013 at 19:27
  • Cloud Compare is good for the analysis of sfm data
    – Rob Lodge
    Commented Nov 14, 2013 at 22:42

3 Answers 3


For looking at points in QGIS you could check out lastools and their las2shp tool. That would turn your 3d point cloud (in las format) into a vector point file. lastools even has a QGIS toolbox now for full integration.

Edit: Check out http://opendronemap.github.io/odm/ as an open source alternative to agisoft.



Let me give you some suggestions:

For SFM workflow you can use VisualSFM (http://ccwu.me/vsfm/). It is free for non-commercial use. Here you can also transform coordinates using GCP or GPS data in EXIF file of images/input gps file.

The output of PMVS can be stored and opened in Meshlab. From there you can export .ply file (I guess should not be saved as binary (uncheck respective checkbox while saving .ply)

You can open the .PLY file with GRASS GIS. Check for extension v.in.ply.

If you use GRASS GIS 7, you can even use v.ply.rectify extension, (http://grass.osgeo.org/grass70/manuals/addons/v.ply.rectify.html) where you also provide GCP file and the result shall be rectified. However I dont know how to deal with setting up coordinate system yet.

From this point it is possible to export as .SHP e.g - v.out.ogr.

Hope this helps :)


I have been working on a similar workflow.

1) Create model in photoscan and save as .obj file
2) Import .obj file into meshlab.
3) Export mesh as xyz file, without normals. (You may wish to alter the orientation of the model first using the manipulator tool to rotate about the z axis)
4) In QGIS, create layer from xyz file
5) From Saga tools, select raster creation tools - Thin plate spline.
6) Specify resolution - depends on data set, but 3000 x 3000 and run. This will create a DEM layer.
7) Once you have created the DEM, you can georeference it and start processing it. For buildings, I have found that a good way of displaying it is to combine a coloured DEM by height with a slope layer to define walls, buildings, roofs etc.

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.